➢ Nondisplaced scaphoid fractures can be effectively treated nonoperatively, with union rates approaching or, in some series, exceeding the rates attained with operative intervention.
➢ The evidence supports equal outcomes when using a short arm or long arm cast for the closed treatment of nondisplaced scaphoid fractures. Also, equivalent outcomes have been demonstrated with or without a thumb spica component to the cast.
➢ Operative intervention is the recommended treatment for displaced scaphoid fractures.
➢ Advanced imaging should be obtained if clinical suspicion is present for a scaphoid fracture with negative radiographs more than 2 weeks after the injury. In some settings, it may even be more cost-effective to obtain advanced imaging sooner.
Background and Epidemiology
As the most commonly fractured bone in the human carpus, the scaphoid accounts for approximately 60% of all carpal fractures1. The incidence of scaphoid fractures has varied in the literature from 22 to 141 per 100,000 person-years1-5. More recently, Van Tassel et al., using the National Electronic Injury Surveillance System database, found the incidence to be considerably lower at 1.47 per 100,000 person-years6. Although the reported incidence has varied, most sources have agreed that scaphoid fractures are more common in younger patients and in male patients2,5-7. Despite their frequency, these fractures can be challenging to diagnose and to treat effectively. Early diagnosis and appropriate treatment are necessary to prevent altered wrist kinematics and arthrosis associated with scaphoid malunions and nonunions8-11.
The scaphoid is the largest bone in the proximal row12, with a complex three-dimensional shape that has been described as a twisted peanut13. It serves as a tie-rod linking the proximal and distal carpal rows14. It has 5 articulations within the wrist, including the distal part of the radius proximally, the lunate on the proximal ulnar side, the capitate via the large ulnar concavity, and the convex distal articular surface, which articulates dorsoulnarly with the trapezoid and radiovolarly with the trapezium15. This complex shape often makes radiographs difficult to interpret and displacement difficult to assess16. Because displacement is associated with nonunion17,18, this difficulty in interpretation has led many to recommend advanced imaging19-21. As a treating surgeon, the unusual shape can make implant placement difficult as well22.
The majority of the scaphoid, approximately 80%, is covered in cartilage providing few entrances for vascularity14. Gelberman and Menon described the vascular supply to the scaphoid in 15 fresh cadaveric specimens. They found that the majority of the blood supply enters the dorsal ridge via a branch of the radial artery. This branch supplies the proximal 70% to 80% of the scaphoid in a retrograde fashion, and the distal 20% to 30% of the scaphoid is supplied by volar branches of the radial artery. Gelberman and Menon suggested a volar approach to the scaphoid to preserve the dorsal vascularity23. Because of its retrograde blood supply, the proximal pole is at higher risk for osteonecrosis and nonunion13,23-26 (Fig. 1).
Biomechanics and Scaphoid Mechanism of Injury and Disease
The scaphoid is an integral component of the intercalated proximal carpal row. The strong connection of the scaphoid to the lunate by the scapholunate interosseous ligament limits scapholunate motion14 and stabilizes the proximal row. This stability is predicated on an intact scaphoid. The primary mechanism of scaphoid fracture is hyperextension of the wrist beyond 95°, which commonly occurs by a fall onto an outstretched hand27,28. An unstable fracture of the scaphoid allows the lunate to extend under load with the triquetrum. The proximal pole of the scaphoid rotates dorsally with the lunate and the distal pole remains flexed via its attachments to the trapezoid and trapezium. This results in the classic humpback deformity associated with displaced scaphoid fractures8 (Fig. 2).
Several classification schemes have been suggested. On the basis of the inclination of the fracture line, Russe classified fractures as horizontal oblique, transverse, or vertical oblique. Transverse and horizontal oblique fractures are the result of compressive forces and are often stable and minimally displaced. Meanwhile, shearing type forces result in vertical fractures, which are almost always unstable and displaced. Vertical fracture lines are relatively rare, representing only 5% of scaphoid fractures29. Similarly, the Mayo classification is based on the anatomic planes of the fracture line, including the proximal, middle, and distal-third scaphoid fractures. The distal-third scaphoid fracture is further classified as a fracture of the distal articular surface or the distal tubercle30. Herbert and Fisher classified fractures on the basis of stability. Type-A fractures are stable patterns including incomplete waist fractures and tubercle fractures. Type-B fractures represent unstable fracture patterns such as distal oblique fractures, complete waist fractures, proximal pole fractures, trans-scaphoid perilunate injuries, and comminuted fractures. Type-C fractures have delayed union. Type-D fractures have nonunion31. Although these classifications schemes can be useful, Desai et al. found only fair intraobserver and interobserver reliability with no correlation with fracture union32.
There have been over 25 described physical examination tests to help identify scaphoid fractures. A high index of suspicion for scaphoid fracture is warranted when physical examination findings include pain with palpation of the volar (distal tuberosity) or dorsal (anatomic snuffbox or proximal pole) aspect of the scaphoid. The axial compression of the thumb metacarpal (scaphoid compression test) may also elicit pain. The sensitivity of these examinations is extremely high, but the specificity is very low individually. When snuffbox tenderness, scaphoid tubercle tenderness, and pain with thumb movement are all found within 24 hours after the injury, the sensitivity is 100% and the specificity is 74%33. Anatomic snuffbox tenderness has been shown to be the most sensitive test for acute fractures of the scaphoid, with rates reported from 87% to 100%34. The axial compression of the thumb metacarpal (scaphoid compression test or longitudinal thumb compression test) may also elicit pain, although it is less sensitive. The variables of male sex, sports injury, scaphoid tubercle tenderness, and anatomic snuffbox pain with wrist ulnar deviation have been shown to be independently significant positive factors for scaphoid fracture and all 4 combined factors are 91% predictive of fracture35,36.
Fracture displacement is crucial in determining fracture stability and risk of nonunion. When history and physical examination are suspicious for fracture, wrist radiographs should be obtained. There is no universally accepted series, although multiple views (typically 4 views) are recommended by most authors. The most common views include neutral wrist posteroanterior, lateral, semipronated oblique, and ulnar deviation7,37. Unfortunately, even with multiple views, radiographs will fail to reveal the fracture in up to 27% of cases34. Scaphoid waist fractures are best seen on an ulnar-deviated posteroanterior view with 20° elbow flexion38, but the lateral view is good for tuberosity and distal-third scaphoid fractures39. The lateral view can also aid in identifying a humpback deformity by assessing the capitolunate, radiolunate, and scapholunate angles. If a fracture is identified, treatment may then ensue primarily on the basis of the presence of displacement.
No consensus exists as to the optimal method of determining displacement or the critical amount of fracture displacement. This limits the ability of the literature to adequately compare displaced and nondisplaced fractures40. Traditionally, radiographic criteria for displacement are defined as a scapholunate angle of >60°41, a 1-mm gap42,43, or a radiolunate angle of >15°44. Computed tomography (CT) is typically considered the best imaging modality for determining displacement and angulation and is recommended when considering operative treatment compared with nonoperative treatment37.
Buijze et al. found radiographs and CT scans to be unreliable in assessing scaphoid fracture stability compared with arthroscopy45. Arthroscopy was used in 44 patients, half of whom had displaced fractures and half of whom had nondisplaced fractures on preoperative radiography. Intraoperative assessment of stability and displacement of the fracture with arthroscopy found a sensitivity of 34%, a specificity of 93%, and an accuracy of 55% for radiographs in determining stability; for CT scans, the assessment found a sensitivity of 62%, a specificity of 87%, and an accuracy of 70%. It still remains uncertain if these fractures are at a higher risk of nonunion compared with other truly stable nondisplaced fractures. Our ability to predict fracture stability is lower than previously thought46.
Diagnosis of some acute scaphoid fractures can be difficult because they may be radiographically occult at initial presentation47. Although some authors have preferred magnetic resonance imaging (MRI) over CT and bone scintigraphy for diagnosis of occult fracture, Yin et al. showed in their meta-analysis that MRI and CT had high sensitivities and specificities; bone scintigraphy was comparable with them in sensitivity but had lower specificity48. Other reviews have shown 96% sensitivity and 89% specificity for bone scintigraphy, 98% sensitivity and 99% specificity for MRI, and 94% sensitivity and 96% specificity for CT19,21. Early MRI in occult fractures has been shown to be cost-effective when analyzing societal cost and may reduce the need for unnecessary immobilization49. Early CT has been suggested to prevent unnecessary immobilization without increased societal cost50. An additional benefit to MRI may be found in subacute or chronic cases in the assessment of the vascularity of the proximal pole of the scaphoid and in the identification of other ligamentous and soft-tissue injuries that may alter the treatment plan32 (Fig. 3).
Nonoperative Compared with Operative
Treatment of acute scaphoid fractures is dependent on many factors, including fracture displacement, anatomic location, comminution, and patient activity or profession. Despite extensive research in the treatment of scaphoid fractures, controversy still exists in determining operative versus nonoperative treatment of various acute scaphoid fractures31,51,52. Operative treatment of displaced scaphoid fractures is more accepted, but operative treatment of nondisplaced waist fractures is more controversial53. Internal fixation of displaced scaphoid fractures is generally accepted because of an unacceptable rate of osteonecrosis, delayed union, and nonunion54.
A randomized controlled trial by Vinnars et al. demonstrated no significant difference in union rates of nonoperatively and operatively treated acute, nondisplaced scaphoid fractures at a median follow-up of 10 years. Union rates were 100% in a long-term follow-up with 35 patients who had undergone nonoperative treatment and 40 patients who had undergone operative treatment55. A recent randomized controlled trial by Clementson et al. demonstrated a 100% union rate of 14 operatively and 21 nonoperatively treated nondisplaced or minimally displaced waist fractures. No difference in grip or pinch strength was identified between the 2 cohorts, but an increased range of motion was noted in the nonoperative group. Operative treatment led to increased rates of radiographic radioscaphoid arthritis at 21% compared with only 9.5% in the nonoperatively managed group and can lead to scaphotrapeziotrapezoidal arthritis if a volar approach is utilized56. In a meta-analysis, Ibrahim et al. found no difference in union rates and an increased rate of complications with operative treatment52.
Nonoperative treatment is usually recommended for nondisplaced distal pole and waist fractures. Union rates of minimally displaced acute scaphoid fractures have been shown in the literature to range from 77% to 100%42,55,57. Dias et al. showed a CT-confirmed union rate of 77% of 44 patients with fractures at 12 weeks with immobilization51, and Buijze et al. showed an overall union rate of 98% of 62 minimally displaced and nondisplaced distal and waist fractures treated with immobilization58.
The preferred method of cast immobilization for scaphoid fractures has varied widely in the literature. Immobilization can include above-the-elbow casts, below-the-elbow casts, various wrist positions, and possible inclusion of a thumb spica. Although earlier literature suggested more rapid time to union with above-the-elbow casting57, in a systematic review and meta-analysis, Doornberg et al. showed no significant differences in union rate, pain, grip strength, time to union, or osteonecrosis for varied methods of nonoperative treatment59. With regard to below-the-elbow casting with or without the thumb inclusion in nondisplaced fractures, Buijze et al. showed no significant difference in treatment types with CT-confirmed union at 10 weeks of 85% of 31 patients who had undergone below-the-elbow casting without thumb inclusion and 70% of 31 patients who had undergone below-the-elbow casting with thumb inclusion58. This confirmed the previous findings of Clay et al. showing no difference in a randomized controlled trial of 392 patients with and without thumb immobilization60. Other studies in the literature have shown no difference in union rates, for either above-the-elbow or below-the-elbow casting, between thumb-spica and non-spica casts57,61.
The definitive determination of the length of immobilization has not been established. Often, immobilization length is determined by follow-up radiographic and clinical examination after 8 to 10 weeks58. Hambidge et al. showed that the removal of immobilization at 8 to 12 weeks led to union of 89% of 121 fractures62. Immobilization may be required up to 12 to 14 weeks for patients at a high risk of nonunion63. Recent studies have utilized the CT-measured extent of union to guide treatment. Immobilization discontinuation has been suggested after >50% of union is reached53,64.
The impetus for operative intervention of scaphoid fractures has included the desire to avoid prolonged periods of immobilization, stiffness, decreased grip strength, and delay in return to work or sport. Anatomic rigid fixation of scaphoid fractures has been shown to lead to faster time to union41,63,65, reduced risk of nonunion51,65, improved functional outcome, and earlier return to work66.
Indications for operative treatment include fracture instability, displacement, angulation, malrotation, and fracture of the proximal pole. Dias et al. found a union rate of 100% of 39 fractures that were treated operatively with screw fixation51. Trumble et al. also achieved a union rate of 100% in 35 acute displaced waist fractures treated with open reduction and internal fixation67.
Operative intervention has been shown by some to lead to a significantly faster time to union65. Bond et al. showed 7 weeks to union for the operative group, whereas the nonoperative group went on to union at 12 weeks41. In contrast to these findings, others have found no difference in time to union68.
Volar or dorsal approaches are commonly used and clinical and biomechanical data exist to support both approaches as safe and effective69,70. Volar approaches are preferred for distal fractures and dorsal approaches are preferred for proximal fractures. Volar approaches are at increased risk of scaphotrapeziotrapezoidal arthritis, although the reported incidences are lower than previously suspected (<3% in one recent study)71. The dorsal approach places the extensor pollicis longus and scaphoid blood supply at a higher risk.
The most commonly used method for fixation of the scaphoid is screw fixation. Early fixation techniques included the use of a solid screw and jig as described by Herbert and Fisher. This technique has been supplanted by cannulated headless compression screw fixation. Cannulated screws are placed through a volar percutaneous, dorsal percutaneous, or mini-open approach14,72. Percutaneous techniques are reserved primarily for nondisplaced fractures or for arthroscopy-assisted techniques in a displaced fracture. Volar and dorsal open techniques are also commonly used for displaced fractures. Studies have supported the placement of screws along the central axis of the scaphoid into subchondral bone. Dodds et al. and McCallister et al. demonstrated biomechanically that a centrally placed screw provides 43% greater stiffness and 39% increased load to failure compared with an eccentrically placed screw73,74. Anthropometric studies have suggested safe screw lengths of 27 mm in men and 23 mm in women, although, more often, 18 to 22-mm screws are utilized. Consideration should also be given to use of a “mini” screw or retrograde insertion secondary to the small width of the proximal pole in women75.
Precise screw placement is critical to minimize iatrogenic injury and to improve biomechanical strength. Obtaining central screw placement can be difficult, and some authors prefer a volar transtrapezial approach to improve accuracy of screw placement76. In most fractures, central screw placement is the preferred technique. Infrequently, in some fracture types, this may not be ideal or practical. Luria et al. demonstrated that screw placement perpendicular to the fracture site placed through the scaphoid tuberosity during a volar approach had similar stability to a screw placed down the central axis. This also avoided placement of the screw across the scaphotrapeziotrapezoidal joint77. In some severe proximal pole fractures, perpendicular screw placement may be preferred to central axis placement to gain good purchase (Fig. 4).
The use of fluoroscopy is advocated to reduce the rate of subchondral bone penetration and screw prominence78. Complication rates of up to 29% have been reported with operative intervention79. In an attempt to avoid complications associated with internal fixation of scaphoid fractures, various techniques have been described. A limited-incision or “mini-open” technique has been advocated instead of percutaneous placement to avoid complications such as screw prominence80. Arthroscopy-assisted techniques have also been described56,81 and have been shown to have acceptable results82. Computer-assisted navigation has also been investigated and was found to have no decrease in overall procedure time, to have reduced fluoroscopy time, and to be only as accurate as traditional methods83. A summary of graded evidence-based treatment recommendations has been included for reference (Table I).
Operative and nonoperative complications are found in treating scaphoid fractures. Delay in diagnosis or missed diagnosis is commonly found. Failure to appreciate fracture instability is also a common cause of inappropriate treatment. Both of these complications can lead to nonunion, osteonecrosis, posttraumatic arthritis, and flexor and extensor tendon ruptures66,84-86. Nonunion rates of scaphoid fractures treated with immobilization are approximately 5% to 10%, with higher rates found in proximal pole fractures. Other risk factors that increase the risk of nonunion are displacement of >1 mm, history of osteonecrosis, vertical oblique fracture pattern, advancing age, and nicotine use. Osteonecrosis has been found to occur in up to 13% to 50% of all scaphoid fractures, with increased incidence with more proximal fractures (Fig. 4). Proximal-fifth scaphoid fractures have been shown to have rates approaching 100% in some studies18,31,42,87. Complications exist with operative treatment and most commonly include aberrant screw placement, osteonecrosis, malunion, nonunion, and need for implant removal (Fig. 5). The rates of complications with a surgical procedure vary from 0% to 29%. The majority of studies have shown low complication rates29. Bushnell et al. found a 21% major complication rate and an 8% minor complication rate with operative treatment79.
Malunion occurs in a predictable manner, with relative extension of the proximal fragment through the intact scapholunate interosseous ligament and flexion of the distal fragment via attachments to the scaphotrapeziotrapezoidal joint. This relationship results in a shortened scaphoid with a humpback deformity, which is radiographically defined as an intrascaphoid angle of >35°. Malunion of the scaphoid results in altered carpal kinematics, which can result in continued wrist pain, stiffness, and arthrosis8,11. Although it appears that mild malunions are well tolerated, the literature is less clear regarding more severe malunions. Amadio et al. found that only 27% of malunions resulted in satisfactory outcomes, but 83% of fractures that healed with normal alignment led to satisfactory results in cases at least 6 months after union8. However, no correlation between degree of malunion and patient outcomes was found in 42 cases at the 1-year follow-up in a study by Forward et al.88. In the only long-term follow-up study comparing well-aligned scaphoid unions and malunions, Jiranek et al. found no difference in patient outcomes at a follow-up of 11 years after the patients had undergone anterior corticocancellous grafting for symptomatic nonunions as an index procedure89. Although there are conflicting data with regard to the effect of malunion on clinical outcome, there is sufficient evidence to support the use of corrective osteotomies in symptomatic malunions90.
The authors acknowledge Bryce A. Van Doren and the OrthoCarolina Research Institute staff for manuscript preparation and submission.
Investigation performed at the OrthoCarolina Hand Center, Charlotte, North Carolina
Disclosure: There was no source of external funding for this study. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work and “yes” to indicate that the author had other relationships or activities that could be perceived to influence, or have the potential to influence, what was written in this work.
- Copyright © 2016 by The Journal of Bone and Joint Surgery, Incorporated