Nonvascularized Bone Grafting for Scaphoid Nonunion

Abstract

Background

Surgical management of scaphoid nonunions requires not only stable fixation but restoration of carpal alignment and reconstruction of bone defects. The latter can be done with either vascularized or non-vascularized bone grafts, depending largely on surgeon preference.

Materials and Methods

This article describes the use of non-vascularized bone grafts for scaphoid nonunions and examines reported outcomes. We also describe the senior author's preferred surgical treatment, the hybrid Russe procedure.

Description of Technique

The hybrid Russe procedure utilizes a corticocancellous strut from the volar aspect of the distal radius to restore anatomy in scaphoid nonunions with flexion deformities. Once the alignment of the scaphoid and associated lunate postural deformities are corrected, fixation then proceeds with a headless compression screw. This combination resulted in healing of 17 scaphoid fracture nonunions at an average time of 15 weeks.

Conclusions

The literature does not demonstrate a difference in union rates when comparing the use of vascularized and non-vascularized grafts for scaphoid nonunions. When the proximal pole of the scaphoid can be salvaged, the choice of fixation is left to the surgeon's discretion.

Failure to recognize a scaphoid fracture or failure of fixation can lead to nonunion and carpal instability with late-stage degenerative changes that occur in a predictable pattern of scaphoid nonunion advanced collapse (SNAC) arthritis. ¹ Common treatment methods include vascularized and nonvascularized bone grafts, with or without cortical bone as needed to address structural defects. Indications for vascularized bone grafts vary by surgeon and institution but are thought to be helpful in cases where the blood supply to the proximal pole is impaired. However, three recent meta-analyses failed to identify differences in union rates with the use of vascularized or nonvascularized grafts. ^{2 3 4} Further, the use of radiographs, bleeding bone from the proximal pole at surgery, or high-resolution magnetic resonance imaging (MRI) was found to have no correlation with the histologic viability of the scaphoid proximal pole as assessed by two independent histopathologists in a prospective series of 35 consecutive patients treated for scaphoid nonunion. In this series, despite histopathological evidence of absent or impaired vascularity in more than 50% of the patients, 33 of 35 patients healed in 12 weeks, and all but one scaphoid nonunion healed with nonvascularized bone graft and rigid internal fixation. ⁵ It is worth emphasizing that despite concerns regarding impaired vascularity of fracture fragments and the possible need for vascularized bone grafts to address this, there is no literature consensus on the best method(s) to predict vascularity or healing. ³ For this reason, there is a paucity of evidence-based conclusions regarding indications for vascularized bone grafting.

Evaluation

History and Physical Examination

History should focus on timing of injury, symptoms, and prior operative and nonoperative treatments. In patients with previously undiagnosed scaphoid fractures, information about prior injury and chronicity of symptoms should be obtained. In those who have failed operative management, assessment of fixation strategies, hardware, and approaches should be performed. Patient factors, including tobacco use, compliance, metabolic bone health, and vitamin D deficiency should also be evaluated. ^{6 7} Physical examination findings, including range of motion and grip strength of the affected and uninjured extremity should be recorded.

Imaging

Posteroanterior, true lateral, and scaphoid views of the wrist should be performed to determine the fracture location, fragment size, proximal pole fragmentation, presence of increased lateral scaphoid angulation or “humpback deformity,” ⁸ degenerative changes, and dorsal intercalated segment instability (DISI). Each of these factors is essential for surgical planning. Advanced imaging is also performed, and often includes both computed tomography (CT) and high-resolution MRI. CT provides precise detail of bony anatomy and bone loss, and if performed with 0.6-mm overlapping slices, can be reformatted in true long axis scaphoid sagittal, true long axis scaphoid coronal, and custom-reformatted planes as needed. ⁹ This helps characterize fracture fragment quality such as fragmentation, osteolysis, bone loss, and fracture line orientation, which is useful for planning trajectory and starting point for the cannulated screw guidewire. ¹⁰ CT also helps identify sclerosis and chronicity, which has been shown to correlate with time to union. ¹¹ In cases of severe deformity, CT of the contralateral wrist is helpful to guide restoration of anatomic alignment, by superimposing three-dimensional segmented models of the injured and opposite sides to determine the size and shape of the resulting wedge defect, the bony loss from prior fixation, and a new path for rigid fixation ( Fig. 1 ). CT is also essential for determining union postoperatively; a scaphoid fracture is generally considered to be united when more than 50% trabecular bridging across the fracture site is demonstrated by CT. ^{5 12}

Fig. 1.

Use of 3D reconstructions of the scaphoid from affected and contralateral side CT scans to determine the size and location of bony defect and new screw trajectory. ( A ) 3D rendering of a left scaphoid nonunion. ( B ) Asymptomatic right scaphoid, mirrored. ( C ) Right scaphoid overlaid over left scaphoid nonunion and realigned. ( D ) Bony defect calculated and shown in orange. ( E ) New screw trajectory in red with prior screw void in gray. CT, computed tomography; 3D, three-dimensional.

MRI has been utilized to evaluate vascularity of the proximal pole, though MRI is highly dependent on the imaging algorithm chosen, slice thickness, resolution, and formatting of the imaging sequences. Hypointensity on T1-weighted imaging and hyperintensity in T2-weighted imaging have been described as indicators of fragment ischemia; however, these findings do not correlate with the presence of punctate bleeding at surgery, ¹³ histologic grading of ischemic necrosis, or likelihood of union. ³ The use of gadolinium-enhanced MRI has been described to evaluate the vascularity of the proximal fracture fragment; however, this also does not correlate with the rate or timing of fracture union. ^{14 15} Importantly, cartilage-sensitive MR sequencing is useful to evaluate loss of cartilage in chronic injuries, the presence and grading of posttraumatic arthritis in cases of SNAC, and the presence of associated carpal ligament injury. ^{16 17}

Operative Techniques

Depending on the location of the nonunion, the surgical approach can be volar or dorsal. Typically, scaphoid waist fractures with humpback deformities are best treated with a volar approach where the fragments can easily be “wedged open” to reduce the deformity and accurately restore anatomic length and alignment. ¹⁸ Cortical bone graft can be harvested as needed from the volar aspect of the distal radius through the same approach. In addition, when DISI is present, defined as a radiolunate angle more than 15 degrees in the dorsal direction, ¹⁹ reduction of lunate extension to neutral not only reduces the DISI deformity but also rotates the attached scaphoid proximal pole into anatomic alignment and reduces the humpback deformity. ²⁰ A temporary transradial Kirschner wire (K-wire) may be used to fix the reduced lunate and simultaneously augment scaphoid fixation by limiting motion of the proximal pole. ^{21 22} In proximal pole nonunions, a dorsal approach allows improved access to the fracture site for debridement and repair. Typically, small proximal pole fragments are not associated with a humpback deformity because of the stabilizing effects of the dorsal ligament complex. ²³ A dorsal approach has not been demonstrated to reduce scaphoid vascularity. ²⁴

Cancellous Autograft

When the amount of bone loss and deformity is minimal, the fracture site is debrided of sclerotic bone and fracture debris down to bleeding bone, and cancellous autograft is placed to fill the defect. Graft volumes of 2.4 to 2.7 mL can be obtained from the dorsal or volar metaphysis of the distal radius via the use of a cortical window. ²⁵ Complication rates from cancellous bone harvest from the distal radius have a rate of 1.7%, and an iatrogenic fracture incidence of 0.2%. ²⁶ Autograft can also be obtained from the iliac crest; however, this is associated with a second surgical site and complications such as donor-site pain and injury to the lateral femoral cutaneous nerve. ²⁷ While Fernandez reported that iliac crest autograft provided improved compressive properties when compared with cancellous graft, ¹⁸ a cadaver study comparing both found comparable biomechanical strength. ²⁸ Studies comparing distal radius and iliac crest autograft in scaphoid nonunion have not shown a difference in union rates, thus the additional morbidity is rarely justified. ^{29 30}

Alternatively, up to 25 mL of cancellous bone can be harvested from the proximal tibial metaphysis, thereby avoiding the complications associated with iliac crest bone grafting. ³¹ Once the nonunion is adequately grafted, fixation is performed at the surgeon's discretion.

Corticocancellous Autograft

In cases where there is scaphoid deformity, the use of cortical bone in addition to cancellous bone grafting provides improved alignment of the fracture fragments. Parameters such as intrascaphoid angle or scaphoid height-to-length ratio (HLR) can be used to calculate the degree of humpback deformity, with scaphoid HLR demonstrated to have high intraobserver and interobserver variabilities. ³² A recent prospective randomized trial comparing tricortical corticocancellous graft with cancellous-only graft in 98 patients who underwent reduction and Herbert screw fixation of a scaphoid waist nonunion found that those with a lateral intrascaphoid angle more than 70 degrees or scaphoid HLR > 0.80 had improved patient-reported outcomes and range of motion with corticocancellous graft. There was no difference in union rates at 24 weeks, with union rates of 94% in patients treated with cancellous graft and 90% with corticocancellous graft. ¹²

Internal Fixation

Once the scaphoid nonunion has been adequately debrided and grafted, internal fixation is performed. Historically, K-wires were used but required prolonged immobilization, and meta-analysis has shown lower rates of healing. ³³ The introduction of the Herbert screw allowed rigid intramedullary fixation with the use of a headless compression screw. ³⁴ Eventually, a variety of cannulated intramedullary screws were developed, along with techniques to align the screw in a more central location within the proximal and distal poles of the scaphoid. ³⁵ Volar fixed-angle plates have also been used and demonstrated a union rate of 96% in 49 patients who underwent scaphoid nonunion repair. Concern with volar plate fixation is impingement on the radius and the increased need for subsequent hardware removal. ^{36 37}

Comparison of fixation methods in a pooled analysis of tricortical wedge grafting showed that intramedullary screw fixation resulted in a 94% rate of union compared with a 77% union rate with K-wire fixation. ³³ A meta-analysis comparing plate versus headless compression screw fixation for scaphoid nonunions showed no significant difference in union rates. Plates were more likely to be used for patients who had failed prior surgical intervention. ³⁸ The use of two headless compression screws with graft has also been described in 19 patients with scaphoid nonunions, resulting in 100% union rates at an average of 3.5 months postoperatively. However, only 11% of patients in this study had a DISI or increased lateral scaphoid angulation, and the authors noted the use of two screws decreased the space available for graft placement. ³⁹

Authors' Preferred Treatment

The hybrid Russe procedure uses a combination of cortical and cancellous techniques to achieve a high rate of union, rapid healing, and correction of deformity. ⁴⁰ The prepped forearm is supinated, and a 4-cm modified Henry approach is extended obliquely across the wrist crease and thenar eminence to the level of the scaphotrapezial joint. The incision is deepened to identify the scaphoid tubercle, and the radioscaphocapitate (RSC) and long radiolunate (LRL) ligaments are divided to expose the fracture site. Care is taken to preserve the volar scapholunate ligament, and the extrinsic ligaments are tagged with nonabsorbable 3–0 sutures for later repair. The scaphoid fragments are then exposed, and a 1.6-mm (0.062 in) K-wire is placed in the distal pole to act as a “joystick” for reduction. The nonunion site is debrided by removing granulation tissue and all sclerotic bone, leaving the dorsal rims of each pole intact to improve leverage. Curettes and the judicious use of a burr, accompanied by saline irrigation, are utilized to cavitate the proximal and distal poles until bleeding bone is reached on each side. The site is packed with a moist sponge and attention directed to the volar metaphysis of the distal radius by dividing the flexor carpi radialis subsheath, retracting the flexor tendons and the median nerve, and lifting a portion of the pronator quadratus. An oval-shaped cortical window, typically measuring 20 × 8 mm, is marked by drilling multiple circumferential drill holes using a 1.1-mm (0.045 in) K-wire. This is carefully lifted off with an osteotome and set aside on the back table ( Fig. 2 ). Abundant cancellous graft is then harvested from the radius and kept in a moist sponge. The lunate is held in neutral alignment with a 1.6-mm transradial K-wire as needed to correct DISI, and the scaphoid distal pole is extended with its joystick to obtain reduction. The resultant distance between the cavitated base of the proximal and distal poles is measured, and a 2- to 3-mm cortical strut is crafted from the oval window and shaped to fit into the defect. A small amount of cancellous graft is placed dorsally prior to insertion of the strut. The two poles are hyperextended with double skin hooks, and the strut is manipulated into place within the scaphoid to maintain the fragments in anatomic alignment. Abundant cancellous autograft is packed around the strut to fill the defect. Fixation is performed with an antegrade or retrograde headless cannulated screw ( Fig. 3 ). The RSC and LRL ligaments are repaired with interrupted nonabsorbable sutures followed by layered closure of the subcutaneous tissues and skin. If utilized, the lunate pin is cut just below the skin. Patients with a lunate wire are immobilized for 10 to 14 days postoperatively in an above elbow splint, continued after suture removal for 4 weeks postoperatively in a Muenster-type cast until pin removal at 4 weeks, and then changed to a short-arm thumb spica cast until 8 weeks postoperatively. Patients are then transitioned into a removable orthosis until CT scan at 10 weeks confirms radiographic union. An example of pre-operative and post-operative imaging is shown in ( Fig. 4 )

Fig. 2.

Intraoperative image showing the design of a cortical strut from the volar aspect of the distal radius. ( A ) After using a K-wire to make drill-holes around an oval-shaped cortical window, an osteotome was used to carefully connect these. ( B ) The cortical window is gently lifted off of the distal radius for later use in the scaphoid fracture nonunion.

Fig. 3.

Illustration describing the use of a hybrid Russe technique for scaphoid nonunions. ( A ) An oval-shaped cortical window is designed in the volar aspect of the radius and a small osteotome is used to gently lift off after perforating the cortex with a K-wire. ( B ) After debridement of the nonunion, the cortical piece is shaped to create a strut that maintains the fracture fragments in anatomic alignment. ( C ) Once adequate alignment is confirmed, cancellous graft is packed into the defect and a headless cannulated screw is inserted across the fracture site, further compacting the graft and ensuring stability. (Source: Reproduced with permission from Lee et al. ⁴⁰ )

Fig. 4.

(A) Pre-operative images of a 17-year-old male patient with a scaphoid nonunion showing a lateral wrist radiograph, AP wrist radiograph and a CT scan in the plane of the scaphoid. The lateral radiograph shows an increased radiolunate angle, indicating DISI. (B) Radiographs and CT scan in the plane of the scaphoid obtained after union of the fracture using the hybrid-Russe technique. The cortical strut is seen dorsal to the screw in the CT scan. (Source: Reproduced with permission from Lee et al. ⁴⁰ )

This combination of the use of a corticocancellous strut from the distal radius, lunate reduction, and a headless compression screw restored carpal alignment in 17 patients; 65% healed at 12 weeks, and all patients healed in a mean time of 15 weeks (range 8–28 weeks). ⁴⁰

Summary and Conclusion

Union rates of nonvascularized bone grafting and internal fixation range from 71 to 100%. ^{40 41 42} Several meta-analyses report an 87 to 88% union rate for nonvascularized bone grafts. ^{2 3} No difference was found in union rates between vascularized and nonvascularized grafts. ^{2 3 43}

Prior literature is insufficient to establish criteria for vascularized bone grafting of scaphoid nonunion, nor improved outcomes of vascularized versus nonvascularized grafts. For all but patients with fragmentation and a nonsalvageable proximal pole, the time-honored principles of adequate fracture debridement, correction of deformity, provision of healthy autogenous bone graft, and rigid internal fixation are required to successfully heal a scaphoid nonunion expeditiously. Creeping substitution, another time-honored principle, enables the well-vascularized distal pole to serve as an engine for trabecular crossing, revascularization, and healing, even in the face of ischemia, absent bleeding bone, or histologic evidence of trabecular necrosis. Whether or not to use one of several vascular graft techniques is largely one of surgeon's preference.

Funding Statement

Funding None declared.