Introduction
Scaphoid malunions and nonunions are common complications of scaphoid fractures. The main risk factors for scaphoid nonunion are delay in diagnosis, fracture displacement, fracture location, and the tenuous blood supply of the proximal fragment with the risk of avascular necrosis (AVN). There are ongoing discussions and controversies regarding the diagnosis and most appropriate treatment. Nonvascularized bone grafts have proven to be effective in most cases. In the setting of AVN, vascularized bone grafting is a recommended treatment.
AVN is difficult to diagnose, with many definitions and different diagnostic methods proposed, including standard plain radiography, MRI, CT, intraoperative bleeding, and histology. , There is conflicting evidence concerning the ability of these methods to accurately identify proximal pole vascularity. A CT-based treatment algorithm can guide bone graft selection for AVN. ,
In the wrist with longstanding scaphoid nonunion, arthritic changes progress. These changes commonly originate between the radial styloid and the distal scaphoid fragment. This is followed by degenerative changes on the scaphoid fossa of the radius, and then in the midcarpal joint. As arthritis occurs, the proximal pole of the capitate migrates toward the scapholunate joint, a process called scaphoid nonunion advanced collapse (SNAC). Degenerative changes and instability increase with time, resulting in osteoarthritis of the wrist in 75% to 97% of cases after 5 years and in 100% after 10 years.
Scaphoid nonunion
Definition and incidence
A scaphoid nonunion is a failure of fracture union 6 months or more after injury. Nonunion occurs in 10% to 15% of acute scaphoid fractures despite appropriate treatment. The major risk factors include delay in diagnosis/treatment, fracture displacement, fracture location, and tenuous blood supply. For dislocated and unstable fractures, some authors report nonunion incidence as high as 55% in conservatively treated fractures. , However, the nonunion incidence varies in different institutes, likely depending on the technical details of such as casting methods and surgical expertise.
Because the major blood supply to the scaphoid is from the radial artery, whose branches enter at the dorsal ridge at level of the waist or distal part and a retrograde blood flow, fracture in the proximal pole (the proximal 20%) may more easily develop AVN. Proximal pole AVN occurs in 13% to 50% of scaphoid pole fractures. Humpback deformity of the scaphoid and dorsal intercalated segment instability (DISI) are often associated with scaphoid waist nonunion and with malunion.
Classification
Nonunions are graded by location, deformity, and vascular status. Given the wide variety of anatomy and vascularity, there is no consensus around a classification system and its prognostic factors. In 1996, Filan and Herbert distinguished type C fractures with delayed union and type D with established nonunions and four subtypes ( Table 9.1 ). The classification scheme of Mack and Lichtman differentiates stable and unstable types. The stable nonunions have a firm fibrous nonunion that prevents deformity of the scaphoid. Unstable nonunions are associated with some degree of carpal collapse deformity. The natural history may lead to the degenerative pattern of SNAC ( Table 9.2 ).
Type | Characteristics |
---|---|
C | Delayed union |
D1 | Fibrous union/stable |
D2 | Pseudarthrosis/unstable |
D3 | Pseudarthrosis/unstable, sclerosis, impaired vascularity |
D4 | Pseudarthrosis/unstable, fragmented proximal pole, avascular necrosis |
Type | Characteristics |
---|---|
I | Stable nonunions, nondisplaced, no degenerative change |
II | Significant displacement or instability dorsal intercalated segment instability, no degenerative change |
III | Early degenerative change, radioscaphoid arthritis with joint space narrowing |
IV | Scaphoid nonunion advanced collapse/arthritis in radioscaphoid and midcarpal joint |
V | Generalized arthritis with radiolunate joint involved |
Slade and Dodds described six grades dependent on the quality of nonunion ( Table 9.3 ). The classification is useful in determining the applicability of minimally invasive procedures. They reported that internal fixation using minimally invasive techniques is sufficient and effective in fractures with delayed presentation, fibrous union, and nonunion without major bone loss (types 1 to 3 and some type 4). Treatment decisions become more complicated in types 5 and 6 and for any of the subtypes.
Grade | Category | Characteristics |
---|---|---|
1 | Delayed presentation | Scaphoid fractures with delayed presentation (4–8 weeks) |
2 | Fibrous nonunion | Intact cartilaginous envelope, minimal fracture line at nonunion interface, and no cyst or sclerosis |
3 | Minimal sclerosis | Bone resorption at nonunion interface <1 mm with minimal sclerosis |
4 | Cyst formation and sclerosis | Bone resorption at nonunion interface <5 mm, cyst formation, and maintained scaphoid alignment |
5 | Cyst formation and sclerosis | Bone resorption at nonunion interface >5 mm and <10 mm, cyst formation, and maintained scaphoid alignment |
6 | Pseudo arthrosis | Separate bone fracture fragments with profound bone resorption at nonunion interface. Gross fragment motion and deformity is often present. |
Subtypes | Category | Associated characteristics |
a | Proximal pole nonunion | Greater risk of delayed or failed union |
b | Avascular necrosis | Fracture must heal and revitalize |
c | Ligamentous injury | Diagnosis by imaging or arthroscopic, direct observation |
d | Deformity | Deformity must be corrected, structural bone graft and rigid fixation |
To our knowledge, these classifications provide certain guidance to treatment selection, but none are satisfactory. Therefore we recommend that these classifications only serve as a general guide rather than a definitive rule for clinical decision making. In other words, an ideal classification does not exist, and a better classification system is necessary.
Clinical presentation and diagnosis
Clinical presentation.
An individual with a scaphoid nonunion commonly presents with long-lasting mild or dull pain on the radial side of the wrist, with pain aggravation during daily use or with wrist weight-bearing. Limited wrist motion is common. However, some patients with nonunion may present with very mild pain or pain that does not interfere with daily use because of the pseudoarticulation that has formed between the two segments of the fractured scaphoid. The most common clinical signs are tenderness in the anatomic snuff box or at the scaphoid tubercle, dorsal swelling, and decreased grip strength. , In patients with a longstanding history of scaphoid nonunion and subsequent SNAC arthritis, pain can be severe and range of motion markedly reduced. These patients typically do not want to use the hand due to pain and loss of wrist functionality. However, some patients do not have any pain.
Diagnostic tools: Nonunion.
The diagnosis of scaphoid nonunion relies on x-rays findings. Clinical presentation alone cannot confirm the diagnosis. Therefore plain radiographs and, often, CT are necessary for patients with a history of acute scaphoid fracture and prolonged clinical problems after treatment or without timely treatment.
Plain radiographs include posteroanterior (PA), lateral, and scaphoid views. Scaphoid-view radiographs can be made as PA radiographs taken with the wrist in ulnar deviation or oblique view with 45% of pronation. Stecher projection with closed fist and ulnar inclination of the wrist may provide additional information.
In plain radiographs, displacement of the fracture segments and humpback deformity (defined as having a lateral intrascaphoid angle of >45 degrees) should be sought. Sclerosis, cystic changes, and areas of bone resorption are often present, but they are not reliable to predict vascularity. These signs of nonunion may not be evident for several months after the initial injury or treatment, although clinical presentation indicates the possibility of development of nonunion. Therefore clinical follow-up and repeat plain radiographs are often necessary.
CT scans provide more detail and more accurate information regarding the nonunion ( Fig. 9.1 ). CT scans may show osseous resorption zones along the fracture line, as well as adjacent pseudocystic inclusions in the scaphoid fragments. CT can help detect these changes earlier and with finer morphologic details. The nonunion cleft often appears larger than what is shown in plain radiographs. Displacement and instability are important risk factors for nonunion of scaphoid fractures. Several measurements in CT scans, such as the lateral intrascaphoid angle or the height to length ratio, may be useful in preoperative planning. ,

Increased density (sclerosis) of the proximal pole and absence of converging trabeculae between the fracture fragments on CT may correlate with intraoperative and histologic evidence of AVN, although some investigators do not find such correlation of CT with histology. The analysis of the complex shape of the scaphoid needs 3D reconstructions, and the uninjured contralateral side can be imaged for comparison. , Three-dimensional CT imaging is helpful to distinguish stable from unstable fractures ( Figs. 9.2 and 9.3 ).


MRI is useful in detecting hidden fractures. MRI was reported to assess the vascularity of the proximal pole, , but it may be unreliable. With an unenhanced MRI, normal bone viability has a high signal intensity. Hypointense areas of bone in T1 and T2 weighted sequences may indicate AVN, but they may also indicate ischemia and even viable bone. , Contrast-enhanced MRI has been found to have higher sensibility in detection of AVN compared with nonenhanced MRI. However, ingrowth of nonspecific inflammatory tissue into the proximal pole from the nonunion site results in contrast enhancement despite the necrotic nature of the bone. ,
Diagnostic tools: AVN.
AVN, occurring in some scaphoid nonunions, is easy to suspect, and yet hard to ascertain. In plain radiographs, remarkably increased density of bone in plain radiographs suggests AVN, but that is uncertain. In MRI, hypointense areas of bone may indicate AVN. It is debatable whether MRI findings accurately diagnose AVN. Therefore MRI findings should be interpreted with great care. It is generally considered that MRI findings cannot indicate viability of the bone. Many colleagues do not rely on MRI to diagnose AVN. Intraoperatively, if the bone does not have punctate bleeding, AVN is likely, which is used popularly. Histologic analysis is reliable, but not feasible intraoperatively.
Treatment
If clinical symptoms are very mild or if a pseudoarticulation has developed with only mild impairment to wrist function, patients often do not wish to have surgery and can be treated conservatively. But the nonunion will not likely heal.
The mainstream treatments are surgical corrections. The goals of surgery are to heal the fracture, correct carpal deformity, and prevent arthritis and carpal instability. To determine which surgical treatment is appropriate, factors such as prior surgical history, size and location of the defect, vascularity, and stability should be considered. Some general guidelines are summarized in Box 9.1 .
- •
Bone grafting is often necessary after curettage of the nonunited site.
- •
Bone graft donor sites have no significant differences in outcomes.
- •
Nonvascularized or vascularized bone grafts produce comparable outcomes.
- •
However, vascularized bone graft (pedicled or free) is more preferable for the patients suspected of avascular necrosis in the proximal segment (proximal pole) and for bringing additional blood supply to help healing.
- •
Depending on the nonunion site, either a palmar or a dorsal approach can be used. Different local donors can be used for a pedicled vascularized bone graft.
- •
Both a compression screw and 2 K-wires are commonly used for fixation and have almost comparable outcomes.
- •
Cancellous bone can be used alone for a small defect (usually 2–5 mm) after curettage of the nonunion site. For a larger defect cortical bone has to be used. Wedge bone grafting is necessary to correct a large defect with scaphoid deformity.
- •
Some patients with scaphoid nonunions are asymptomatic with good function because a pseudoarticulation develops, for whom surgery is not needed.
Internal fixation without bone grafting.
The most common surgery is open reduction, correction of humpback deformity with bone graft, and stable fixation. In minimally displaced or nondisplaced scaphoid nonunions, percutaneous screw fixation alone without bone grafting can be attempted ( Fig. 9.4 ). Rigid fixation without bone grafting can lead to bone healing in some patients. Hegazy et al reported 21 cases of scaphoid waist nonunions treated with percutaneous compression screw fixation without bone graft. Inclusion criteria were an intact cartilaginous envelope, minimal fracture line, no cyst or sclerosis, and no humpback deformity. However, internal fixation alone is not a commonly used method for treating scaphoid nonunions.

A minimally invasive method of internal fixation with bone graft is an arthroscopically assisted percutaneous internal fixation with a headless compression screw inserted with a dorsal percutaneous technique. , This minimally invasive technique may prevent the complications of an open surgery and gives minimal surgical trauma to the blood supply and soft tissues. , However, this approach is technically demanding, and there are no comparative studies showing this approach is superior to the open approach. Therefore this approach remains unpopular.
Surgical approaches: Volar or dorsal.
Most surgeons prefer internal fixation through a volar or dorsal incision of approximately 4 to 5 cm. The volar approach preserves the dorsal blood supply and provides access to middle and distal third fractures with access to correct a humpback deformity. The dorsal approach provides improved exposure of the proximal pole with easier internal fixation placement, but it can disrupt the tenuous vascular supply. ,
Debridement at the fracture site and bone grafting.
The nonunion site is accessed, and a thorough curettage is performed to debride the nonunion site. The soft fibrous union should be excised. The sclerotic surfaces of the two fracture segments should be removed until soft, healthy cancellous bone is exposed on both segments. The amount debrided will vary, but usually about 2 to 3 mm of such sclerotic bone should be removed to expose the healthy bony surface.
A variety of donor sites can be used, which is subject to personal preference. There is no conclusive evidence to indicate which donor site is best. The commonly used bone graft is cancellous bones from the distal radius, allograft of cancellous bone, vascularized bone graft from the distal radius (pedicled), or free vascularized bone graft from the ilium or knee area. See the text below for discussion of choice of donor sites and evidence to support the use of each. Nevertheless, the ideal donor site is unknown, and there is a question regarding the need of vascularized bone graft.
Choices of internal fixations: Compression screw versus kirschner wires.
After grafted bone is placed between the two fractured segments, rigid internal fixation is important ( Fig. 9.3 ). Headless compression screws and Kirschner (K) wires are commonly used methods of internal fixation. The screw is more commonly used. Some surgeons consider that K-wires alone may not ensure immobilization due to a lack of compression ( Fig. 9.5 ). However, using K-wires with bone grafting is preferred by some surgeons (including Jin Bo Tang), as they consider that compression does not favor nutrition including the blood supply and there is no basis for compression to the grafting bone. A systematic review in 2015 included 48 papers reporting on 1,602 patients with scaphoid nonunion. They were divided into subgroups according to different graft donors with screws (33 subgroups) or K-wire (27 subgroups) fixation. The authors did not find a significant difference, with union incidence of 88% for screws and 92% for K-wires, although patients with screw fixation were allowed to mobilize earlier. Among these patients, there were 125 proximal pole fractures with nonunions and AVN, among which 60 fractures were treated with screws for the grafts. The union incidence was 81% using screws and 97% using K-wires. Union incidence was encouragingly high given the traditional views that avascularity of the proximal pole is a poor prognostic indicator for scaphoid nonunion, although this view has been disputed. It appears that rigid fixation of the graft is more important than compression in treating scaphoid nonunion with a graft. It appears preferable to use K-wires in treating nonunions of the proximal pole with a graft because the proximal pole is small and may be fragmented during compression with screws, compressing the graft, and the proximal pole may further deprive synovial nutrition to the proximal pole and graft. Previous views on screws and compression were based largely on a metaanalysis two decades ago, which showed an improved union incidence of 94% with screw fixation compared with K-wire fixation of 77%. Despite attempts to advance the surgical management of scaphoid nonunion, approximately 10% persist after treatment.

Placement of a screw within the central third of the longitudinal axis of the scaphoid improves the stability of the fracture fixation. Preoperative 3D CT imaging may help precise screw placement ( Fig. 9.6 ). In selected cases, palmar angular stable plate fixation is an option. Leixnering et al reported a union of all 11 patients with scaphoid waist nonunions treated with plate fixation. Two of the authors (Rohit Arora and Gernot Schmidle) currently use such plating in cases of failed previous osteosynthesis, instability, bone defect, or short distal fragments ( Fig. 9.7 ).


Bone grafting sources and methods: Nonvascularized
Nonvascularized bone grafts are the conventional treatment and have been effective in most situations.
Distal radius versus iliac crest bone graft.
The common donor sites are distal radius and iliac crest. The union incidence and functional results of the techniques are equivalent. , The iliac crest graft has more donor site morbidity but seems to possess superior osteogenic properties compared to distal radius grafts. ,
Corticocancellous versus cancellous graft.
An unstable scaphoid waist nonunion with humpback deformity is commonly treated with structural corticocancellous bone grafting and internal fixation. Cancellous bone grafting was originally used to treat stable nonunions without deformity. However, cancellous bone graft with rigid fixation shows comparable results with shorter interval to union. If the proximal and distal fragments are large enough, rigid screw or K-wire fixation can function as a strut, and cancellous bone grafting can be an option to treat unstable scaphoid nonunions. Sayegh and Strauch compared results of corticocancellous and cancellous grafts and found comparable healing incidence. Two of the authors (Rohit Arora, Gernot Schmidle) prefer plate fixation after corticocancellous bone graft because the proximal pole is often quite short and only a few screw threads can engage healthy tissue, jeopardizing stability ( Fig. 9.8 ). Additionally, the palmar plate works as a buttress against the humpback deformity.

The Matti-Russe procedure is the traditional technique and has recorded union incidence, but it has been associated with a high incidence of failure in cases of diminished or absence of punctate bleeding at surgery. With this technique, it is difficult to correct a humpback deformity. Therefore the Matti-Russe procedure is not commonly used nowadays.
A better method of graft placement is the Fisk-Fernandez technique, a method with good reported outcomes and union incidence. , The defect in the scaphoid waist or proximal pole is filled with a triangular or trapezoidal wedge-shaped bone graft, commonly obtained from the iliac crest. , This donor site is favored for patients with scaphoid nonunion and AVN of the proximal pole. , Matsuki et al treated 11 patients with an iliac bone graft and Herbert screw fixation for proximal pole nonunion, among which two patients had AVN. Healing was achieved in all patients, regardless of the vascularity of the proximal pole. In reporting a study of the Scaphoid Nonunion Consortium in 2018, Rancy et al followed 35 consecutive patients with scaphoid nonunions in a prospective longitudinal registry. All nonunions were treated with curettage, nonvascularized autogenous grafting, and headless screw fixation. Thirty-three of the 35 scaphoids healed by 12 weeks, with 30 having focal or robust remodeling activity and a 94% incidence of healing. Healing was defined as ≥50% bony bridging on CT images in the plane of the scaphoid. Among the 33 patients, nine patients had ischemia of the proximal pole but no infarction in preoperative MRI images, and 28 patients had impaired vascularity assessed intraoperatively. The authors concluded that proximal pole infarction is decidedly rare and that vascularized bone grafting is seldom required for a successful outcome with proximal pole nonunions.
Bone grafting sources and methods: Vascularized
Over the past two decades, vascularized bone grafting has gained popularity in treating scaphoid nonunions of the proximal pole, particularly those affected by AVN and for patients with longstanding nonunion or who have failed previous surgery. , , The exact value of using a vascularized bone grafting is uncertain. , , Controversy remains as to the need for vascularized grafting in the context of AVN, and a systematic review in 2015 found no significant difference in union incidence between the two graft methods.
There are two types of vascularized bone grafting for the nonunion: local pedicled and free vascular grafts.
Pedicled bone from dorsal distal radius.
Pedicled vascularized bone grafts from the distal end of the radius have many potential benefits. The corticocancellous bone can replace the deficient bone stock at the nonunion site and provide structural support ( Fig. 9.9 ). , This is a popular method used by many hand surgeons because the graft is easy to harvest and only one surgical field is required for grafting and internally fixing the scaphoid fracture. The surgical outcomes are generally satisfactory.

Fibrous tissue and any avascular bone in the nonunion site are curetted. Most often, the graft from the distal radius is harvested based on 1,2-intercompartmental supraretinacular artery (1,2-ICSRA). This artery is dissected through a dorsal curvilinear incision to expose the first and second extensor compartments. The 1,2-ICSRA vessel overlies the extensor retinaculum between the two compartments. The periosteum with cortical bone and underlying cancellous bone are harvested with preservation of the artery pedicle and a 5-mm cuff around the vessel to allow venous return. The size of the bone to harvest depends on the defect in scaphoid to fill, and additional nonvascularized distal radius cancellous bone graft is used when necessary. After surgery, the patient is placed into a long-arm splint or cast, followed by casting and/or splinting until radiographic union. The 2,3-intercompartmental supraretinacular artery (2,3-ICSRA) pedicled bone graft can be used with similar surgical techniques, but the use of 2,3-ICSRA is less popular than 1,2-ICSRA.
In the presence of dorsal intercalated segmental instability (DISI) or AVN, this procedure may not work well. Chang et al reported union at an average of 16 weeks after surgery in 34 of 48 (71%) scaphoid nonunions treated between 1994 and 2003. Twenty-seven of the fractures were stabilized with cannulated screws, 15 with K-wires, four with both screws and K-wires, and three without internal fixation. The choice of internal fixation was made by the surgeon based on fragment size and stability. Failures in achieving union occurred in 9 of 14 (64%) fractures with DISI and/or humpback deformity. In proximal scaphoid fractures with proximal pole AVN, 5 of 8 fractures with screw fixation and 2 of 4 with K-wire fixation had union; the overall union incidence was 50%. Chang et al and other surgeons consider that the 1,2-ICSRA is not a suitable option for the treatment of scaphoid nonunion with DISI or with AVN of the proximal pole.
Although these grafts are usually placed as dorsal inlay grafts, they also have been successfully used as trapezoidal volar intercalated grafts. Henry reported union in all 15 patients at a mean of 11.5 weeks after surgery. All patients had humpback deformity and proximal pole AVN as diagnosed by MRI and intraoperative inspection.
Vascularized volar distal radius graft.
A pedicled graft from the volar aspect of the distal end of the radius has been described by Kuhlmann et al. With this technique, a structural interposition graft to correct the humpback deformity can be harvested. Grafts from the volar ulnar side of the distal radius have higher structural stability and have successfully been used to correct deformity ( Fig. 9.10 ). Dialiana et al and Gras and Mathoulin have reported favorable outcomes in union, correction of carpal malalignment, and restoration of scaphoid length. ,

Free vascularized iliac crest graft.
Pechlaner et al used a free vascularized bone graft from the iliac crest based on its vascular pedicle and anastomosed to the radial artery ( Fig. 9.11 ). Gabl et al reported union in 12 of 15 patients with proximal pole AVN after this graft. The lead author and colleagues treated 21 patients who had failed prior procedures or MRI evidence of AVN, which was confirmed intraoperatively; 16 patients (76%) achieved union. Harpf et al reported 60 cases of scaphoid nonunion (21 with AVN) treated by this graft with union incidence of 92%.

Free vascularized medial femoral condyle graft.
The medial femoral condyle (MFC) can treat the scaphoid nonunion in the setting of AVN and correct the humpback deformity and carpal collapse ( Fig. 9.12 ). The MFC graft is corticocancellous and uses either the longitudinal branch of the descending geniculate artery and vein or the superomedial genicular vessels. The lead author and colleagues evaluated the use of a MFC graft with and without cartilage in 37 patients with recalcitrant scaphoid nonunion with a specific focus on union rates and proximal pole destruction. Our results showed a union rate of 95% and good functional outcomes. Doi et al used vascularized MFCs as volar inlay grafts and achieved union in all 10 patients with AVN of the proximal pole at an average of 12 weeks after surgery. Jones et al reported successful use of the free vascularized MFC graft in scaphoid nonunion with AVN and humpback deformity with failed previous surgery with union in all cases.


Another study compared the use of MFC graft as a structural interposition graft and the use of distal radius pedicled vascularized graft (1,2-ICSRA) in cases of scaphoid nonunion with proximal pole AVN and humpback deformity. All 12 patients treated with the MFC graft healed and had significantly improved postoperative carpal angles. In contrast, in 10 patients treated with the 1,2-ICSRA, only four patients gained union but failed to show improved carpal alignment. Especially in longstanding nonunion with failed screw fixation, the lead author prefers to use a vascularized MFC graft. In these cases, the proximal pole already has bone loss because of the implanted screw. Additionally, the blood supply is compromised by the surgical approach.
Vascularized osteochondral medial femoral graft.
An osteochondral MFC graft includes articular cartilage from the medial femoral trochlea ( Fig. 9.13 ). It can be used for a variety of indications and procedures, including replacement of the entire proximal part of the scaphoid and successful reconstruction and restoration of carpal alignment.

Studies directly comparing vascularized and conventional grafts are scarce.
Although there are many studies focusing on either vascularized or nonvascularized bone graft as the treatment, studies directly comparing vascularized and conventional grafts are scarce. In 2004, Munk and Larsen reported a metaanalysis of 147 articles on 5,249 scaphoid nonunions. Eighty-four percent of the patients treated with nonvascularized bone grafting with internal fixation achieved union with an average immobilization of seven weeks, and 91% of those treated with vascularized bone grafting with or without internal fixation healed at 10 weeks.
In 2016, Ferguson et al reported a systematic review examining 144 articles with 5,464 scaphoid nonunions. They reported an 84% union incidence for vascularized bone grafting (904 patients) and 80% for nonvascularized bone grafting (3,971 patients). In the presence of AVN (543 patients), the incidence of union was 74% for vascularized bone grafting and 62% for nonvascularized bone grafting. AVN was diagnosed in 40 studies in several ways, including radiography, MRI, and punctate bleeding. Union incidence varied considerably. They found the need for high-level evidence.
Ribak et al prospectively randomized 86 scaphoid nonunion patients into groups treated with either distal radius vascularized bone grafting or nonvascularized bone grafting with K-wire fixation. Union was achieved in 89% in the vascularized graft group and 73% in the nonvascularized graft group, a statistically significant difference. Proximal pole AVN was assessed by intraoperative punctate bleeding. In the AVN subgroup, union incidence was 83% with vascularized bone grafting contrasted to 55% with nonvascularized bone grafting.
In the institute of one of us (Ruby Grewal), a clinical trial randomized patients to receive a pedicled vascularized bone graft (28 patients) based on the 1,2-ICSA or a nonvascularized iliac crest graft (22 patients). All had K-wire fixation. Thirty-eight patients were available for union assessment and 23 for clinical measurements. No significant differences were found in union incidence, time to union, complications, patient-reported outcome scores, or wrist mobility and grip strength between the groups. Smokers were 60% less likely to achieve union. When controlling for smoking, patients receiving a vascularized graft were 72% more likely to achieve union.
Additional donor sites and surgical methods.
Corticocancellous bone from the olecranon can be easily accessed through a longitudinal incision in the same surgical field that is used to treat scaphoid injury. Bone substitutes may also be used, although the effectiveness in treating scaphoid nonunion has not been assessed. Vascularized corticocancellous graft from the radius (about 1 × 1 cm) attaching to a capsular flap of the dorsal wrist capsule (of about 2 cm long) is another method. , The graft is harvested from the distal radius ulnar to Lister’s tubercle about 2 cm proximal to the articular surface and can include the dorsal ridge of the distal radius. After fixation of the scaphoid with a cannulated screw, the graft is press-fit into the scaphoid nonunion site. Vascularized bone graft from the base of the second metacarpal or third metacarpal base with a dorsal carpal capsule strip can also be used.
The value of an arthroscopically assisted approach
Arthroscopy helps assess the fracture site for any fragment mobility, firm callus between the fracture fragments, gapping, and the width and depth of the gaps. In 2006, Slade and Dodds reported the technique of percutaneous bone grafting, with a cancellous bone graft harvested from the distal radius and fixed with a screw. They classified scaphoid fractures from delayed presentation and fibrous nonunion to pseudarthrosis with bone deformity ( Table 9.3 ); however, this is truly a documentation of a variety of pathologies seen with the nonunion, which is too complicated as a practical classification. Minimally invasive internal fixation was shown to be effective in treating fractures with delayed presentation, fibrous union, and nonunion without major bone loss.
Wong and Ho published the result of 69 patients with symptomatic scaphoid nonunion treated with arthroscopic bone grafting. The incidence of healing was 84% to 95%, depending on the vascularity of the proximal fragments. Wong and Ho reported outcomes in 124 patients and showed an overall union incidence of 90% and good clinical outcomes.
Arthroscopic management of scaphoid nonunions provides a surgical method with less damage to the blood vessels and nerves, faster recovery, and the opportunity to diagnose and treat concomitant injuries. The lead author and colleagues prefer this method in cases with no humpback deformity and especially in cases of proximal pole fractures. During the arthroscopy, we assess the mobility and the resorption zone of the fracture site. In cases with no mobility of the proximal pole and no resorption zone, we restore stability with percutaneous screw fixation (antegrade, retrograde) without any bone graft.
An arthroscopic approach can be used for cases with humpback deformity with DISI, in the experience of one of the authors (Toshiyasu Nakamura). The Linscheid correction technique with K-wire fixation between the radius and lunate can correct the DISI. Then, resection of the distal surface of the nonunion of the scaphoid is possible from the midcarpal approach. After resection of the fibrous tissue between the nonunion, the proximal and distal fragments of the scaphoid are fixed with three to four K-wires under fluoroscopic control without traction. Placement of the K-wires are as palmar as possible. The forearm is set back into a traction tower, and arthroscopic bone grafting from the midcarpal approach is possible.
Proximal pole replacement
Osteochondral autograft transplantation.
If the proximal pole is fragmented or necrotic, an osteochondral graft from the MFC offers a favorable solution for proximal pole replacement. Weber et al reported eight patients who underwent osteochondral autograft transplantation (OAT) harvested from the medial or lateral trochlea at an average of 18 months after injury. Four patients had failed prior attempts at scaphoid union surgery, one of whom failed two prior surgeries. Four had no prior surgery. In the procedure, a plug was taken from the medial femoral trochlea, offering a hyaline cartilage surface. The underlying subchondral bone was harvested with the cartilage, which should assist in healing. The plug was pressed to fit into the scaphoid without hardware. The average follow-up was 12 months. All eight OATs healed. CT scan confirmed union in six patients between 6 and 10 weeks. OAT is an attractive procedure for patients with proximal pole scaphoid nonunions associated with an intact scapholunate ligament. This procedure mitigates the need for vascularized bone grafting.
Vascularized medial femoral osteocartilaginous flap has demonstrated successful outcomes in treating scaphoid proximal pole nonunion with a high union rate and restoration of functional range of motion and grip strength. , This procedure, however, is lengthy and technically demanding and may not improve the union rate compared with nonvascularized options. Nonvascularized autograft arthroplasty, such as rib or hamate autograft, has demonstrated encouraging outcomes but still requires prolonged postoperative immobilization and a return to the operating room at around 8 weeks for the removal of K-wires.
Prosthetic replacement of the scaphoid proximal pole with an adaptive proximal scaphoid implant has been shown to prevent SNAC and to have high clinical satisfaction and fast return to sport and work. However, failure was reported in 12%, with persistent pain and/or loss of grip strength. Dislocation was reported in 4%, and long-term outcomes are lacking. This technique does not correct carpal collapse and is indicated only when the proximal pole can be excised without injuring the scapholunate ligament. It should be known that OAT does not provide the structure necessary to correct and resist carpal collapse. This procedure is to restore cartilage alone. In cases where bone and cartilage reconstruction is necessary, where there is scaphoid collapse or scapholunate ligament injury, or where the proximal fragment is too large to excise without compromising scapholunate stability, an alternative procedure is indicated. This could be a vascularized bone graft, a proximal hamate autograft, or a rib osteochondral graft as described below.
Proximal hamate autograft.
The proximal hamate autograft is a relatively new option. The ipsilateral hamate is harvested as an osteochondral graft in the size of the defect to be replaced together with the adhering portions of the volar capitohamate ligament. The proximal pole of the hamate can also serve as a replacement arthroplasty in the setting of proximal pole scaphoid nonunions with collapse, bone loss, and/or osteonecrosis. This graft may address shortcomings of other graft choices by providing a local structural autograft solution with minimal donor site morbidity and with the opportunity to correct carpal collapse and reconstruct the scapholunate ligament without the need of microvascular anastomosis.
Rib osteochondral grafts.
Rib osteochondral grafts have also been described. , Descriptions vary regarding the harvesting site, and fixation is carried out with K-wires, as screws tend to split the graft during insertion. Good results have been reported, but these procedures are burdened with an unfavorable donor site. In the experience of one of the authors (Toshiyasu Nakamura), the donor site problem for this procedure is quite minimal. Veitch et al reviewed 14 patients with deficiency of the proximal pole of the scaphoid who were treated by rib osteochondral replacement arthroplasty. Improvement in wrist function occurred in all except one patient. The authors reported that this procedure can restore the mechanical integrity of the proximal pole of the scaphoid and maintain wrist movement while avoiding the potential complications of alternative replacement arthroplasty techniques and the problems associated with vascularized grafts and salvage techniques. Interested colleagues should go to the cited references for surgical methods and tips from the surgeons who have performed the procedures.
Postoperative care for the above procedures
The patients should be casted for 8 to 10 weeks depending on the bone graft used. After cast removal, hand therapy follows. Plain radiographs and, if needed, CT are used to assess the healing status of the grafted bone at week 4 or 6, and then again later. In general, the cast is needed for at least 6 weeks, commonly 8 weeks. If bony union is not found, casting can be prolonged to 10 or 12 weeks. If partial union is found, a splint can be used after weeks 8 to 10. There is no need to remove the compression screw unless it causes pain. The K-wire is normally pulled out 8 to 12 weeks after surgery when union is confirmed by plain radiograph or CT. CT is particularly useful to judge the healing when there is doubt in findings of the plain radiographs. At the time of removal of cast or K-wire, the union may only partially connect the two fracture segments, which is sufficient for removal of the cast or K-wire, because union of the entire interface may be slow or never occur in some patients.
In hand therapy, aggressive motion should be avoided because it may break the union. Only gentle motion exercise should be performed. Weight-bearing-related hand exercise should be avoided as this can risk damaging the union. After 3 to 4 months, therapy can be more aggressive, and patients may resume normal hand use.
Clinical difficulties, tips, and unsolved problems
The importance of stability and vascularity for scaphoid nonunions is widely agreed upon, but there is an ongoing discussion regarding the most appropriate method of diagnosis and treatment. There is a wide range of opinions on whether to use vascularized or nonvascularized grafting for the treatment. Avascular bone grafts show a high union incidence as long as sufficient vascularity of the fragments is present. Currently, we see that failures often occur in proximal pole fractures with critical vascularization or extensive bone loss, especially with AVN, longstanding scaphoid nonunion, or failed previous surgery. It appears that in these cases vascularized bone graft should be used.
There is a lack of large prospective studies, and retrospective case series and cohorts comprise most of the evidence. The studies use different criteria to define nonunions, and they often do not consider the impact of confounders and use different types of bone grafts according to nonunion characteristics such as site deformity or vascularity. Thus the studies come to different conclusions about the success of bone graft surgery, which makes it difficult to establish evidence-based treatment recommendations.
The presence of AVN of the proximal fragment is a critical issue in the outcome of surgical treatment of scaphoid nonunion and the choice of an appropriate treatment. Therefore diagnosing vascularity is crucial in the choice of an appropriate scaphoid nonunion treatment.
Few definitions have been agreed on in the literature with respect to AVN, and a multitude of options have been described for diagnosing AVN of the proximal scaphoid, including standard radiography, CT, MRI, intraoperative vascularity, and histopathologic assessment ( Box 9.2 ). When examining the literature, one must look carefully at how vascularity was defined and assessed to judge the effectiveness of different bone grafts.
