➢ Evaluation of scaphoid nonunions should include an assessment of proximal pole vascularity and scaphoid malalignment.
➢ Rigid fixation with a headless compression screw is preferable to fixation with Kirschner wires when fragment size allows.
➢ When osteonecrosis is present, vascularized bone graft is preferable to nonvascularized graft.
➢ When collapse is present, scaphoid geometry should be restored with structural graft or cancellous graft and rigid fixation.
➢ Salvage procedures are recommended when scaphoid nonunion-associated carpal collapse and arthritis (scaphoid nonunion advanced collapse) is more extensive than changes at the radial styloid process alone.
Scaphoid fractures occur most commonly through the scaphoid waist1. Fracture initiation and propagation have been postulated to result from hyperextension2, hyperflexion, or axial load3. In one series involving eighty-two fractures of the scaphoid waist, the rate of nonunion was 12%4. The risk of nonunion increases with delayed diagnosis5, inadequate immobilization6, fracture instability7, fracture displacement8, and associated ligamentous injury9,10. As the scaphoid is largely covered with articular cartilage, its vascular supply is precarious, particularly in the proximal-third region, which relies primarily on retrograde flow from the dorsal scaphoid branch of the radial artery11. Thus, proximal-third fractures are especially predisposed to nonunion and osteonecrosis11.
The natural history of symptomatic scaphoid nonunion is a progressive pattern of carpal instability, collapse, and subsequent degenerative change. Sequential radiographic changes occur over time, progressing from sclerosis, cysts, and osseous resorption at the nonunion site; to radioscaphoid joint-space narrowing with beaking of the radial styloid process; and then to capitolunate joint-space narrowing12. The radiolunate articulation is spared13. These degenerative changes occur because of rotary subluxation of the distal scaphoid fragment, resulting in scaphoid nonunion advanced collapse (Fig. 1)14. The pattern of arthritis in cases of scaphoid nonunion is similar to that of scapholunate advanced collapse resulting from disruption of the scapholunate interosseous ligament.
Scaphoid nonunions initially may be minimally symptomatic, with eventual dorsoradial wrist pain that is aggravated at the extremes of motion and with gripping activities. Physical examination may demonstrate radial-sided swelling, snuffbox tenderness, limited range of motion, and diminished grip strength15. Performance of the scaphoid shift test will result in pain16. Although physical examination findings have demonstrated a sensitivity of 100% and a specificity of 74% to 80% for the detection of acute and subacute scaphoid fractures, classic examination findings may be absent in cases of established nonunions15,17.
Radiographs, including posteroanterior, lateral, and scaphoid views, are useful for evaluating the location of the nonunion, the presence of carpal collapse, and the presence of static carpal instability. Scaphoid-view radiographs can be made as posteroanterior radiographs with the wrist positioned in ulnar deviation or as oblique radiographs with 45° of pronation; the former view elongates the scaphoid18,19. Markers of unstable nonunion include displacement of >1 mm19, malalignment or so-called humpback deformity (with a lateral intrascaphoid angle of >45°)20, and static carpal instability such as dorsal intercalated segment instability (with a radiolunate angle of >15°)19,21. Degenerative changes associated with scaphoid nonunion advanced collapse may be apparent at the radial styloid process and/or the distal pole of the scaphoid or midcarpal joints14.
Advanced imaging studies provide additional information that may be useful for surgical planning. Computed tomography (CT) scans showing increased proximal fragment radiodensity are strongly associated with a histological diagnosis of osteonecrosis22. CT scans most clearly demonstrate the extent of bone loss, the degree of displacement, and humpback deformity of the scaphoid20,23,24.
Magnetic resonance imaging (MRI) has been used to diagnose osteonecrosis, which initially was believed to be demonstrated by hypointense bone signal on T1 and T2-weighted sequences25,26. In fact, however, these parameters alone are, at best, 36% to 62% sensitive for detecting osteonecrosis of the scaphoid27,28. Experimental comparisons of quantitative canine carpal bone blood flow, measured by radioactive microsphere entrapment with T1 and T2-weighted magnetic resonance images, have shown no direct relationship between osteonecrosis and MRI findings29. The use of contrast-enhanced MRI may improve test accuracy considerably, although spurious contrast enhancement may occur with ingrowth of nonosseous fibrovascular tissue27,30.
As such, CT scans are integral to our evaluation of scaphoid nonunions and our preoperative planning for their treatment. Non-contrast-enhanced MRI is reserved for assessing the cartilage at the proximal pole of the scaphoid and the scaphoid fossa.
While early descriptions of surgical treatment and bone-grafting for scaphoid nonunion did not include the use of internal or percutaneous fixation31,32, there is general consensus today that rigid fixation is optimal for the successful treatment of scaphoid nonunions33-42. Both Kirschner wires and screws have been used in conjunction with bone-grafting for the treatment of scaphoid nonunion. Stark et al. reported a union rate of 97% (147 of 151) in a study of scaphoid nonunions that were treated with cancellous or corticocancellous grafts and stabilized with multiple Kirschner wires33. Subsequent studies involving Kirschner-wire fixation demonstrated union rates ranging from 56% (fifteen of twenty-seven)34 to 100% (twenty-six of twenty-six)35.
In 1984, Herbert and Fisher described the use of a headless compression screw for the treatment of scaphoid fractures and nonunions36. Since then, numerous retrospective studies have evaluated the use of headless compression screws and, more recently, cannulated headless compression screws, to stabilize scaphoid nonunions37-42. Time to union is significantly decreased if the screw is placed within the central third of the longitudinal axis of the scaphoid (p < 0.05)21, and such placement results in increased stability biomechanically43,44. In the study by Trumble et al., placement of the screw within the central third was achieved in a significantly higher percentage of patients who were managed with cannulated screws compared with those who were managed with non-cannulated screws (94% [seventeen of eighteen] compared with 35% [six of seventeen]; p < 0.01)21. While a finite element analysis model suggested decreased motion at the site of oblique fractures when a screw was placed eccentrically perpendicular to the fracture plane45, a biomechanical study demonstrated no difference in stability between a central screw and a perpendicular screw46. A meta-analysis of the literature demonstrated a significantly improved union rate when screw fixation was compared with Kirschner-wire fixation (94% [sixty-eight of seventy-two] compared with 77% [forty-one of fifty-three]; p < 0.01)47.
Other, less commonly used techniques for the fixation of scaphoid nonunions also have been described. Carpentier et al. reported that iliac-crest bone-grafting and staple fixation was associated with a union rate of 95% (thirty-six of thirty-eight)48; Bumbasirevic et al. reported that the use of an Ilizarov external fixator to provide sequential distraction followed by compression, without bone-grafting, was associated with a union rate of 100% (fifteen of fifteen)49; and Leixnering et al. reported that the use of plate fixation for the treatment of scaphoid-waist nonunions was associated with a union rate of 100% (eleven of eleven)50. While rigid fixation of scaphoid nonunions is often combined with some form of bone-grafting procedure, percutaneous and arthroscopy-assisted internal fixation without bone-grafting also has been reported for the treatment of stable, established scaphoid nonunions with preserved carpal alignment and proximal pole vascularity51-53.
Successful treatment of scaphoid nonunions that are associated with bone resorption relies on meticulous technique and the use of autogenous bone-grafting. Essential surgical principles include correction of scaphoid malalignment, debridement of fibrous tissue, removal of necrotic bone to expose healthy cancellous bone surfaces, addition of bone graft (cancellous or corticocancellous), and stabilization of the scaphoid with internal fixation and/or an intrinsically stable corticocancellous bone graft54.
Cancellous bone-grafting and rigid fixation is the mainstay of treatment for stable scaphoid nonunions without displacement, carpal instability, or osteonecrosis. With the technique initially described by Otto Russe in 1960, the nonunion site is excavated through a volar approach and cancellous iliac-crest bone graft is placed into the nonunion site, followed by immobilization in a long arm thumb-spica cast for twelve to sixteen weeks32. Although Otto Russe reported a union rate of 91% (twenty of twenty-two) without any form of scaphoid fixation, cancellous bone-grafting in isolation is not routinely performed today. In the study by Yasuda et al., all twenty-eight scaphoid nonunions that were treated with bone graft from the distal part of the radius and Kirschner-wire fixation healed at a mean of seven weeks55. In the study by Finsen et al., 90% of thirty-nine scaphoid nonunions healed after treatment with cancellous bone graft from the iliac crest (fourteen patients) or the distal part of the radius (twenty-five patients) and Kirschner-wire fixation, including all fourteen of the proximal-third nonunions56. However, preoperative and postoperative scaphoid alignment was not examined.
With the advent of a cannulated headless compression screw and the benefits of rigid fixation, selected scaphoid waist nonunions associated with a humpback deformity can be successfully treated without the use of structural bone grafts. Cohen et al. observed a 100% union rate, confirmed with CT scans, in a study of twelve scaphoid waist nonunions with a humpback deformity (mean preoperative lateral intrascaphoid angle [and standard deviation], 49° ± 4°) and no osteonecrosis that were treated with an open Kirschner-wire-assisted reduction, fixation with a headless compression screw, and nonstructural cancellous bone graft from the distal end of the radius57. At a mean of forty-one months of follow-up, there were nine excellent and three good results according to the modified Mayo wrist score20, with maintenance of normal scaphoid alignment (mean postoperative lateral intrascaphoid angle, 32° ± 2°). The authors proposed that if adequate purchase can be achieved in the distal and proximal fragments, the headless compression screw could afford structural support. They cautioned that compression generated with insertion of the headless compression screw could cause shortening and malalignment that may contribute to unsatisfactory results. Amadio et al. reviewed the records on forty-six scaphoid fractures and noted that, among the twenty-six scaphoids that had a humpback deformity at the time of union, only 27% (seven) had a satisfactory clinical outcome whereas 54% (fourteen) had posttraumatic arthritis20. Similarly, Jiranek et al. noted that, at an average of eleven years postoperatively, the objective scores for patients who had a humpback deformity were significantly worse than those for patients who had restoration of scaphoid alignment (p < 0.0001)58.
While cancellous bone-grafting for the treatment of scaphoid nonunions is most commonly performed as an open technique, arthroscopy-assisted fixation and cancellous bone-grafting also have been described51,59,60. Slade et al. described an arthroscopy-assisted dorsal approach for headless compression screw fixation of scaphoid nonunions with minimal sclerosis, an intact fibrous shell surrounding the nonunion site, and normal scaphoid alignment51. The authors reported a union rate of 100% (fifteen of fifteen), confirmed with CT scans at a mean of fourteen weeks postoperatively, without any complications. The outcome was excellent for twelve patients and good for three according to the modified Mayo wrist score. Saint-Cyr et al. reported union of five of five scaphoid waist nonunions that were treated in a similar manner60. This technique was later modified to include arthroscopic cancellous bone-grafting from the iliac crest with use of an 8-gauge bone-biopsy cannula, which can be used to advance the bone graft into the predrilled headless compression screw track without complete debridement of the nonunion59. More recently, arthroscopy-assisted debridement and cancellous bone-grafting has been described even for nonunions that previously were treated with standard open approaches61. Wong and Ho reported that this method was associated with a union rate of 91% (sixty-two of sixty-eight), with a mean radiographic time to union of twelve weeks61. Advocates of arthroscopic treatment of scaphoid nonunions attest to the ability to thoroughly assess the nonunion, to diagnose other intra-articular abnormalities, to minimize surgical trauma to ligamentous and vascular structures, and to maintain a favorable biological environment for fracture union61. Key principles in the arthroscopic treatment of scaphoid nonunions include reconstitution of vascular channels with guidewire placement and reaming or predrilling, assessment of proximal pole vascularity, debridement of devitalized bone, preservation of fibrocartilaginous tissue surrounding the nonunion site, correction of fracture displacement and scaphoid malalignment, percutaneous implantation of cancellous bone graft, and augmentation with rigid internal fixation59.
Use of Bone-Graft Substitute
While autogenous bone-grafting is the so-called gold standard, use of allogeneic bone or bone-graft substitutes eliminates donor-site morbidity and pain62,63. Chu and Shih used arthroscopy-assisted reduction, headless compression screw fixation, and injection of bone-graft substitutes (beta-tricalcium phosphate and calcium sulfate) to treat scaphoid nonunions without osteonecrosis, humpback deformity, severe carpal collapse, or arthrosis64. Union was noted in 93% (fourteen) of fifteen scaphoids at a mean of fifteen weeks, with ten excellent and five good outcomes according to the modified Mayo wrist score64.
As bone is resorbed from the site of a fracture nonunion, the scaphoid tends to collapse and shorten. Fisk and subsequently Fernandez modified the Matti-Russe volar technique and recommended the use of internal fixation after placement of an iliac-crest graft as a trapezoidal wedge to correct the scaphoid shortening and carpal malalignment, and Green reported results with the technique65-67. Nakamura et al. evaluated the results of iliac-crest volar wedge grafting and headless compression screw fixation for the treatment of scaphoid nonunions and found that a duration of nonunion of more than five years, proximal pole nonunion, proximal pole sclerosis, and failure to correct the scaphoid deformity were the primary determinants of less-satisfactory results68. Tsuyuguchi et al. similarly demonstrated a significant relationship between the postoperative wrist score and the postoperative scapholunate angle (p < 0.001)69. Union rates following the treatment of scaphoid nonunions with volar autogenous corticocancellous bone grafts have been reported to range from 66% (forty-five of sixty-eight) to 96% (twenty-five of twenty-six)70. The source of autograft (either the iliac crest or the distal part of the radius) has been shown to make no difference in union rates70,71. A meta-analysis demonstrated that nonvascularized autografts do not function as well when osteonecrosis is present, with union being achieved in only 47% of thirty patients who received a nonvascularized graft as compared with 88% of thirty-four patients who received a vascularized graft47.
Arteriovenous Bundle Implantation
Implantation of a vascular bundle in conjunction with bone-grafting can restore vascularity in cases of scaphoid nonunions that are associated with proximal pole osteonecrosis. Based on the technique described by Hori et al., transplantation of an artery and vein into bone can result in the formation of new bone with active proliferation of new blood vessels72. The benefit of this technique is its simplicity, a short operative time, the lack of a need for complex vascular dissection or microvascular surgery, and the ease of manipulating a vascular bundle alone, without an attached bone graft73. Fernandez and Eggli reported successful healing of ten of eleven proximal-third scaphoid nonunions that were treated with a corticocancellous inlay graft from the iliac crest combined with implantation of a second dorsal metacarpal arteriovenous bundle73. At a mean of five years of follow-up, the Mayo wrist scores were excellent for three patients, good for three patients, fair for three patients, and poor for two patients with persistent pain despite healing. Successful healing in the presence of osteonecrosis was attributed to scaphoid stability provided by the iliac crest wedge graft and improved vascularity provided by the transplanted vascular bundle73. In another series, a variety of arteriovenous bundles that included the second dorsal metacarpal artery were transplanted into the sites of eight scaphoid nonunions that were associated with MRI findings suggestive of osteonecrosis74. The resulting 100% union rate and the postoperative MRI findings that were indicative of normalized circulation demonstrated the merit of this method.
When osteonecrosis is present, the use of a vascularized bone graft should be considered to facilitate revascularization of the proximal pole and healing of the scaphoid nonunion. Transplantation of viable osteocytes is postulated to accelerate graft-host bone union through primary bone-healing without creeping substitution or resorption75-77. In addition, Sunagawa et al. reported that the use of vascularized bone grafts enabled revascularization and resulted in active bone remodeling at the sites of previously necrotic proximal poles in a canine model78. Both pedicled and free microvascular transferred bone grafts have been used for the treatment of scaphoid nonunion79-82.
Pedicled Vascularized Bone Graft
Dorsal Aspect of the Distal End of the Radius: Pedicled vascularized bone grafts from the dorsal aspect of the distal end of the radius have many potential benefits when used for the treatment of scaphoid nonunions and osteonecrosis. The corticocancellous bone of the dorsal aspect of the distal end of the radius can replace the deficient bone stock at the nonunion site while providing some measure of structural support75,76. In a canine experimental model, pedicled vascularized bone grafts from the distal end of the radius demonstrated a fourfold increase in bone blood flow over the first two weeks following surgery as measured by radioactive microsphere entrapment83. First described by Zaidemberg et al. and further anatomically defined by Sheetz et al., bone graft from the dorsal aspect of the distal end of the radius based on the 1,2-intercompartmental supraretinacular artery has been commonly employed for the treatment of scaphoid nonunions79,80. Correction of scaphoid and carpal malalignment is of paramount importance in the treatment of any scaphoid nonunion. Although early reports of pedicled grafts based on the 1,2-supraretinacular artery demonstrated 100% union79,80,84, a subsequent study by Chang et al. demonstrated a union rate of only 71%85. Many of these patients had a scaphoid humpback deformity and carpal collapse that were not adequately corrected and were a significant factor in the failure of treatment. They found that only 50% (twelve) of twenty-four scaphoid nonunions with proximal pole osteonecrosis united, and 50% (seven) of fourteen failures were in patients with concomitant osteonecrosis and humpback deformity of the scaphoid and/or dorsal intercalated segment instability. The authors advised against the use of pedicled grafts from the dorsal aspect of the distal end of the radius when both osteonecrosis and scaphoid collapse are present. Low healing rates with the pedicled graft based on the 1,2-supraretinacular artery have been reported by others as well86,87. While these grafts are usually placed as dorsal inlay grafts, they also have been used successfully as trapezoidal volar intercalated grafts, with Henry reporting a union rate of 100% in a study of fifteen patients88. Sotereanos et al. reported that the use of a capsular-based pedicled graft from the dorsal aspect of the distal end of the radius, supplied by the fourth extensor compartment artery, resulted in union of ten of thirteen scaphoid nonunions (including eight of ten that were associated with osteonecrosis) when the graft was placed as a dorsal inlay graft89. Two of the ten patients with osteonecrosis had had mild dorsal intercalated segment instability, which was unchanged postoperatively.
Volar Aspect of the Distal End of the Radius: A pedicled graft based on the palmar carpal artery from the volar aspect of the distal end of the radius initially was described by Kuhlman and colleagues90.With this technique, a trapezoidal graft can be harvested and used as a structural interposition graft to correct a humpback deformity. Two recent studies demonstrated successful outcomes with regard to union, correction of carpal malalignment, and scaphoid length91,92. Although none of the 111 scaphoid nonunions in the report by Gras and Mathoulin were associated with proximal pole osteonecrosis, and sixty-four scaphoids had no instability or malalignment, an overall union rate of 96% was reported for the entire group after treatment with a pedicled bone graft based on the transverse volar carpal artery with Kirschner-wire or headless compression screw fixation92. The study group included forty-two patients with dorsal intercalated segment instability and five with concomitant radioscaphoid arthritis.
Thumb Metacarpal: A bone graft from the thumb metacarpal, pedicled on the first dorsal metacarpal artery, also has been used as a volar interposition graft to treat persistent scaphoid nonunions93. Bertelli et al. reported that this type of graft was used successfully for the treatment of nine of ten scaphoid nonunions; six of the nine were associated with dorsal intercalated segment instability93.
Free Vascularized Bone Graft
Iliac-Crest Graft: In addition to pedicled grafts about the wrist, several free vascularized grafts have been described to address scaphoid nonunions. Gabl and colleagues used free vascularized tricortical grafts from the iliac crest, based on a deep circumflex iliac vascular bundle, for patients with scaphoid nonunion and proximal pole osteonecrosis94. Their most recent study demonstrated union in 76% (sixteen) of twenty-one patients, seven of whom had a humpback deformity95.
Medial Femoral Condylar Graft: To our knowledge, the use of vascularized bone and periosteum from the medial femoral condyle was first described by Hertel and Masquelet96, who suggested its use as a pedicled corticoperiosteal flap based on the articular branch of the descending genicular artery and vein or the superomedial genicular vessels for the treatment of lower-extremity fracture nonunions and segmental tibial bone loss as well as for revascularization in cases of osteonecrosis about the knee. This technique evolved into the use of a free vascularized corticoperiosteal flap as well as a vascularized bone graft for a variety of indications as described by Sakai et al.97. Doi et al. reported that the use of a volar inlay graft for the treatment of scaphoid nonunion was successful in ten of ten patients98. The Mayo Clinic group has reported on the use of medial femoral condylar graft as a structural interposition graft combined with a headless compression screw to restore scaphoid length and carpal alignment in cases of nonunions complicated by both proximal pole osteonecrosis and humpback deformity81,82. In the study by Jones et al., all twelve patients who were managed with this type of graft had union and significantly improved postoperative carpal angles81. However, in a similar cohort of ten patients, the use of a 1,2-intercompartmental supraretinacular artery graft failed to correct carpal malalignment and resulted in a 40% union rate81. The authors stressed that the free medial femoral condylar structural graft can only be utilized if the proximal pole of the scaphoid has an intact cartilaginous shell, is of adequate size for fixation, and is without fragmentation81,82.
Osteochondral Medial Femoral Trochlear Graft: Recently, the free medial femoral condylar vascularized bone graft has evolved to include articular cartilage from the medial femoral trochlea. This type of graft has been used for a variety of indications and procedures, including replacement of the entire proximal part of the scaphoid. Anatomic studies have shown that the transverse branch of the descending genicular artery provides a rich network of periosteal vessels supplying the trochlea of the medial patellofemoral joint99,100. A medial femoral condylar bone graft, including a portion of the medial femoral trochlear cartilage, may be harvested in such a way that the cartilage surface closely approximates the normal proximal scaphoid radiocarpal radius of curvature, resulting in successful salvage of a fragmented proximal pole nonunion101. Bürger et al., in a series of sixteen patients with proximal pole nonunions, reported that fifteen patients (94%) had union without evidence of carpal instability after a mean duration of follow-up of fourteen months102.
Wrist denervation is a useful palliative surgical option for patients who have a painful scaphoid nonunion with carpal collapse and degenerative arthritis103,104. Complete denervation attempts to transect all articular branches supplying the radiocarpal and midcarpal joints105. Braga-Silva et al., in a review of forty-nine wrist denervations, reported 80% improvement in terms of pain when the preoperative levels were compared with the levels after a mean duration of follow-up of three years106. Partial denervation avoids multiple incisions by transecting the terminal branches of the posterior and anterior interosseous nerves through a single dorsal incision107,108. Relief of wrist pain following partial denervation may be similar to that following a formal complete denervation, and partial denervation of the wrist provides results similar to those of a formal complete denervation and does not impair the function of the pronator quadratus109. Wrists with painful scaphoid nonunions that are not amenable to fixation are candidates for either complete or partial denervation as neither procedure precludes future salvage procedures.
Excision of the Distal Fragment of the Scaphoid
Excision of the distal fragment at the site of a chronic nonunion of the distal third of the scaphoid may provide considerable pain relief and improved range of motion if there is minimal associated arthritis110. Malerich et al. reported complete pain relief in 68% (thirteen) of nineteen patients at forty-nine months after excision, with substantial gains in flexion and extension as well as in deviation110. Preoperative capitolunate arthritis was found to result in further degenerative change and was associated with two unsatisfactory results. Certainly, partial excision of the scaphoid can result in carpal collapse with dorsal intercalated segment instability111, as was found to be the case in nearly half of such patients in one report112. Lunates with a hamate articular facet (type II) have been associated with more radioulnar motion and a smaller radius of curvature of the head of the capitate compared with those with an absent hamate facet (type I)113, and thus they may be at risk for progressive capitolunate arthritis after excision of the distal fragment of the scaphoid as a result of altered carpal kinematics113,114.
Carpal collapse resulting from scaphoid nonunion results in arthrosis over time. When localized to the radial styloid process and/or scaphoid articulation, radial styloidectomy may serve as a primary or adjunctive technique115,116. Both open and arthroscopic methods have been used117, most commonly in conjunction with another procedure, including scaphotrapeziotrapezoid arthrodesis118, four-corner arthrodesis119, and proximal row carpectomy120. An overzealous resection risks increasing radial, ulnar, and palmar translation of the carpus if more than 3 to 4 mm is resected because of the loss of the palmar radioscaphocapitate ligament121. Ruch et al., in a study of three patients, reported pain relief, patient satisfaction, and satisfactory objective results according to the modified Mayo wrist score after short-term follow-up (mean, thirty months)122. The authors proposed that, with appropriate technique and cautious selection of patients, this technique causes minimal morbidity, relieves mechanical pain, and improves motion.
Scaphoidectomy and Four-Corner Arthrodesis
The progressive pattern of arthritic changes resulting from scapholunate and scaphoid nonunion advanced collapse was first described by Watson and Ballet, who termed the condition scaphoid nonunion and advanced collapse (or SNAC)123. Their use of scaphoid excision and four-corner arthrodesis remains a reliable method for the treatment of advanced scaphoid nonunion advanced collapse arthritis today. El-Mowafi et al. studied the cases of ten patients with stage-II (radiocarpal) or stage-III (radiocarpal and midcarpal) scaphoid nonunion advanced collapse arthritis who underwent scaphoidectomy and four-corner arthrodesis. The authors reported seven good results, two fair results, and one poor result based on the modified Mayo wrist score124. Krakauer et al. reported a union rate of 91% (twenty-one of twenty-three) and a mean flexion-extension arc of 54° after scaphoidectomy and four-corner arthrodesis for wrists with stage-II and stage-III scapholunate advanced collapse125. Four-corner arthrodesis carries the risk of nonunion, implant-related issues, dorsal impingement (with dorsal circular plates), and loss of motion secondary to a malpositioned lunate126. Patient satisfaction and pain relief have been reported to be better after scaphoidectomy and four-corner arthrodesis than after proximal row carpectomy, particularly in younger patients and in manual laborers127-130. Dacho et al. noted higher mean postoperative grip strength in patients with stage-II and stage-III scapholunate advanced collapse, expressed as a percentage of the grip strength of the contralateral hand, in patients who underwent scaphoidectomy and four-corner arthrodesis (72%) than in those who underwent proximal row carpectomy (50%)131.
Proximal Row Carpectomy
Scaphoid nonunions with a preserved lunate fossa and minimal midcarpal arthritis (stage-II scaphoid nonunion advanced collapse lesions) may be amenable to proximal row carpectomy. Advocates of proximal row carpectomy have cited a more rapid recovery, absence of implant-related complications, no risk of pseudarthrosis, and a less technically demanding procedure as benefits over scaphoidectomy and four-corner arthrodesis132. Proximal row carpectomy may result in improved range of motion133; better Disabilities of the Arm, Shoulder and Hand (DASH) scores134; and fewer complications135 in properly selected patients. It should not be performed in patients with stage-III scaphoid nonunion advanced collapse, as midcarpal arthritis may lead to persistent pain133. In addition, most patients without initial degenerative changes in the capitate will develop radiocapitate articulation arthritis over time, although it may be minimally symptomatic. Techniques such as osteochondral resurfacing of the capitate136 and dorsal capsular interposition with137 or without138 resection of the capitate head may extend indications when capitate cartilage is damaged. However, caution is needed when selecting proximal row carpectomy over other salvage operations for the management of younger patients. In a recent study, conversion of proximal row carpectomy to total wrist arthrodesis was required for 18% (four) of twenty-two patients at a mean of seven years after the carpectomy; all of these conversions were in patients who were thirty-five years of age or younger at the time of the proximal row carpectomy139.
Total Wrist Arthrodesis
Total wrist arthrodesis is the final salvage procedure for the treatment of scaphoid nonunion. It is most commonly used in patients with advanced arthritis when motion-sparing options are not appropriate or have failed. The procedure remains a valid primary treatment of scapholunate advanced collapse arthritis, particularly for manual laborers who perform load-bearing work140. When total wrist arthrodesis is used after the failure of other procedures, pain relief is often incomplete but is still of value to patients125. In one series, twenty patients with scapholunate advanced collapse and/or scaphoid nonunion advanced collapse that had previously been treated with scaphoidectomy and four-corner arthrodesis underwent conversion to total wrist arthrodesis for the treatment of progressive radiolunate arthritis, ulnar translocation of the carpus, infection, nonunion, or unremitting pain141. Although only 35% (seven) of twenty wrists were pain-free at rest and 10% (two) of twenty wrists were pain-free with strenuous use, 90% (eighteen) of the twenty patients were satisfied with the outcome of the total wrist arthrodesis.
Critical Analysis of the Literature
There are many limitations in the establishment of an evidence-based medicine algorithm for the treatment of scaphoid nonunions. Several classification schemes have been proposed along with treatment recommendations, yet they are not all-encompassing with regard to nonunion characteristics and treatment options51,59,128. In addition, most of the literature comprises case series (Level-IV evidence) with diverse patient populations and medical histories, including different therapeutic and/or surgical procedures, prior surgical interventions, scaphoid fracture locations, and time intervals following injury. In many cases, important variables such as alterations of the intrascaphoid angle, carpal collapse, the degree of radiocarpal and/or midcarpal degenerative change, associated ligamentous injuries, and the presence of osteonecrosis are not clearly described. Also, the definition of successful scaphoid union among these studies is heterogeneous, with no standardized method to assess for union clinically or radiographically; of note, the authors routinely made CT scans to determine the degree of bone healing at six weeks postoperatively. Nonetheless, recommendations for the care of scaphoid nonunions, based on the presence or absence of osteonecrosis (as defined by the presence or absence of visible punctate bleeding from the proximal pole at the time of surgery)67 and scaphoid malalignment (humpback deformity) with minimal to no degenerative change (up to and including stage-I scaphoid nonunion advanced collapse if a concomitant radial styloidectomy is performed), are listed in Table I.
The evaluation and treatment of scaphoid nonunions pose many difficult problems for the hand surgeon. A thorough assessment of the fracture location, the degree of bone loss, the presence of osteonecrosis, evidence of degenerative change, and scaphoid malalignment (or humpback deformity) is imperative for successful outcomes. Although many different methods of fracture fixation and bone-grafting techniques have been described, several underlying principles are apparent, including the use of rigid fixation when possible, correction of scaphoid length and malalignment, bone-grafting to encourage union and to restore mechanical stability, and improvement of proximal pole blood supply when osteonecrosis is present. Fragmentation of proximal pole fractures has been an indication for a salvage procedure; however, free vascularized osteocartilaginous bone-grafting may prove a reasonable alternative. Symptomatic scaphoid nonunion advanced collapse arthritis, when present, cannot be ignored. Adjunctive use of radial styloidectomy, or selection of another salvage procedure, is required in these instances to provide predictable results.
Source of Funding: No external funding was received in the preparation of this manuscript.
Investigation performed at the Division of Hand Surgery, Department of Orthopedic Surgery, Mayo Clinic, Rochester, Minnesota
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. One or more of the authors, or his or her institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. No author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
- Copyright © 2014 by The Journal of Bone and Joint Surgery, Incorporated