Hand Surgery
1st Edition

21
Fractures of the Carpus: Scaphoid Fractures
Scott W. Wolfe
Fractures of the scaphoid are common and affect predominantly young and productive individuals. The highly mobile scaphoid plays a critical role in coordinating the kinematics of the two carpal rows and is subject to considerable compressive, rotational, and shear forces during hand and wrist motion. In addition, the scaphoid has a peculiar and limited retrograde blood supply that is vulnerable to disruption by trauma; consequently, scaphoid fractures are prone to a high rate of delayed union and nonunion. Untreated fractures of the scaphoid predictably progress to nonunion, and the dissociation of the two scaphoid fragments leads to a relentless progression of abnormal carpal kinematics and degenerative change. Prompt diagnosis and treatment have been facilitated by improved imaging techniques for detection of occult fractures. A trend toward more aggressive surgical management of scaphoid fractures with rigid internal fixation has resulted in higher rates of union and an earlier return to sports and activities in selected individuals. This chapter focuses on diagnostic principles, assessment of stability and prognosis for healing, and treatment methods directed at reducing the potentially high morbidity of this difficult injury.
DEMOGRAPHICS
The annual incidence of carpal fractures in the United States was reported at more than 678,000 in 1995, of which at least 70% (475,000) represent fractures of the scaphoid (1,2 and 3). Fractures of the scaphoid are second only to fractures of the distal radius among fractures of the upper extremity but have a decidedly different demographic profile. Unlike distal radius fractures, scaphoid fractures are more common among men and have a peak incidence in the second and third decades (4). Average time for healing of a nondisplaced scaphoid fracture in a cast is 9 to 12 weeks (1,5,6,7,8,9,10 and 11), accounting for a considerable degree of lost time and productivity in this young and active population. Missed or delayed diagnoses are not uncommon and may result in additional morbidity from secondary changes, including nonunion, collapse deformity, and degenerative arthritis (12,13).
Scaphoid fractures in children are uncommon, and tubercle fractures are more frequent in children than adults (14,15,16 and 17). Most children’s fractures occur from low-energy falls and heal readily with cast immobilization at an average of 7 to 8 weeks (15,17,18,19,20,21 and 22). Pediatric scaphoid fractures are not immune to complications, however, including nonunion and avascular necrosis (Fig. 1) (15,16 and 17,23).
SURGICAL ANATOMY
The scaphoid is an unusually shaped bone, likened to both a peanut (24) and a boat (from Greek skaphos = boat) but resembling neither upon close scrutiny (Fig. 2). Analogous to the head of the talus or the head of the femur, the scaphoid is almost entirely covered in cartilage, with the exception of a thin dorsal strip of its waist. This reduces its capacity for periosteal healing and increases its propensity to progress to delayed union or nonunion after fracture (25,26). The scaphoid’s largest articular face is its convex proximal surface, which articulates with the scaphoid facet of the distal radius. A large concave sulcus opposite the radial articular facet acts as a socket for the radial aspect of the nearly spherical capitate head. The thinner waist portion contains the nonarticular dorsal ridge, along which the critical dorsal ridge vessels traverse and that acts as an insertion point for both the dorsal component of the scapholunate interosseous ligament (27), as well as the dorsal intercarpal ligament (28). The distal pole is pronated, flexed, and ulnarly angulated with respect to the proximal pole and presents separate articular surfaces to the trapezium and trapezoid distally. A distinct interfacet ridge on the scaphoid separates the trapezial and trapezoidal articular surfaces and may constrain the axis of rotation of these joints to a single oblique plane (29). Its hemispherical
P.382

tuberosity projects palmarward, serving to anchor the scaphotrapezial and transverse carpal ligaments and as an important entry point for scaphoid vascular supply.
FIGURE 1. Untreated scaphoid fracture in an adolescent that progressed to nonunion with attendant shortening and early degenerative changes.
Early studies had identified three distinct vascular pedicles to the scaphoid on the distal, lateral-volar, and dorsal margins (30), but more recent studies suggest that there are just two major vascular conduits (8,11,31). All would agree, however, that the blood supply of the scaphoid proximal pole is tenuous, dependent almost exclusively on intraosseous vasculature derived from perforators along its dorsal ridge. Using vascular injection and clearing via the Spalteholz technique, two predominant vascular pedicles have been identified: one entering the scaphoid tubercle and supplying its distal 20% to 30% and the other arising from the radial artery and comprising the dorsal ridge vessels (Fig. 3) (25,31,32). The dorsal ridge vessels originate from the radial artery and penetrate via several foramina to supply the proximal 70% to 80% of the scaphoid. Obletz and Halbstein studied 297 dried cadaveric scaphoid preparations and showed that 13% of specimens had no vascular perforators proximal to the waist and another 20% had only a single perforator proximal to the waist (33). In a study by Gelberman et al., the majority of the vascular foramina were positioned at or distal to the scaphoid waist, and no intraosseous anastomoses were identified between the two arterial systems in any specimen (32). Because of its unusual retrograde vascular supply, the scaphoid has the greatest risk among the carpal bones for avascular necrosis after fracture. Temporary interruption of the blood supply to the proximal fragment is a virtual certainty with proximal pole fractures (8,24), but if rigidly immobilized, the proximal pole still has the capacity to heal and revascularize through trabeculation (11,34). Radiographic changes of avascular necrosis develop in 25% to 40% of proximal pole fractures (9,10 and 11,33), making delayed union and nonunion far more common among this fracture subtype.
Because of a theoretical concern for interruption of the dominant dorsal ridge vascular supply, many authors have cautioned against a dorsal operative approach to scaphoid fractures. Several clinical series using the dorsal approach, however, have demonstrated excellent healing rates and have not borne out a predicted increase in avascular necrosis or proximal pole collapse (35,36 and 37). As blood supply through the scapholunate interosseous and radioscapholunate ligaments is negligible, it is likely that the bone has an excellent capacity for revascularization. The bone’s reliance
P.383

on intraosseous flow, however, necessitates absolute immobilization to allow revascularization to occur, and in proximal third fractures, this is most readily accomplished with rigid internal fixation. In recent years, a trend toward early and rigid fixation of proximal third fractures has greatly reduced nonunion rates in this location (36,38,39).
FIGURE 2. Computer-generated three-dimensional reconstructions of scaphoid anatomy. A: Dorsal view, demonstrating the concave capitate fossa to the right and convex radial articular surface proximally. B: Ulnar view into capitate fossa, with tubercle to the right. (Courtesy of Joseph J. Crisco III, Ph.D., with permission.)
FIGURE 3. Schematic representation of vascular supply from the radial artery. (Redrawn from Gelberman RH, Panagis JS, Taleisnik, J, et al. The arterial anatomy of the human carpus: part I. The extraosseous vascularity. J Hand Surg [Am] 1983;8:367–376, with permission.)
BIOMECHANICAL CONSIDERATIONS
Although the scaphoid has been traditionally included as a member of the proximal carpal row, it was argued by Gilford et al. that the scaphoid is better characterized as a member of both the proximal and distal carpal rows (40). From a kinematic perspective, the scaphoid is an independent bone whose three-dimensional motions are dependent on the direction and degree of hand positioning in space (41). Through its stout proximal and distal ligamentous connections, the scaphoid serves to coordinate and smooth the motions of the proximal and distal rows and has been likened to a slider-crank mechanism that stabilizes an inherently unstable dual linkage system (Fig. 4) (40,42). The kinematic effect of an unstable scaphoid fracture is a dissociation of the proximal and distal carpal rows that permits the natural tendency of the two carpal rows to fail by collapsing. This has been shown experimentally by Smith et al. (43) and is demonstrated clinically by the collapse pattern seen with chronic scaphoid nonunion, a condition coined scaphoid nonunion advanced collapse (44,45). Under axial load, the two halves of the scaphoid collapse into a flexed or humpback posture (Fig. 5) (43,46,47). The proximal half, untethered by its ligamentous connections to the distal carpal row, moves synchronously with the lunate. When not balanced by an intact scaphoid, the lunate follows its natural propensity to rotate into extension, an effect caused by a dorsal pole that is narrower than its palmar pole and the unchecked tendency of the triquetrum to rotate into extension on the oblique proximal articular surface of the hamate. The capitate flexes at the midcarpal joint and translates dorsally with the conjoined hamate on the proximal carpal row. This collapse pattern, or dorsal intercalated segment instability, leads to a predictable sequence of degenerative disease due to abnormal mechanics and altered load distribution. Arthritic change arises first at the radial styloid articulation with the distal scaphoid pole, followed by degeneration of the midcarpal joint, and ultimately to pancarpal arthritis (Fig. 6) (48).
Scaphoid waist fractures have been consistently replicated in cadaveric experimental protocols that simulate a fall on the outstretched wrist. Weber and Chao applied
P.384

load to wrists in different degrees of extension and determined that the scaphoid waist fails with the wrist between 95 and 100 degrees of extension with load applied to the radial surface of the palm (49). In this position, the authors demonstrated that the radial collateral ligament complex became lax and hypothesized that this structure normally exerts a protective effect against fracture. With the proximal pole trapped between the scaphoid fossa of the radius and the tense palmar extrinsic ligaments, the full force of the applied load on the scaphoid distal pole was concentrated at its waist, and nine of 10 specimens failed, with a clean fracture at the mid-waist (Fig. 7).
FIGURE 4. The stabilizing effect of the scaphoid on the inherently unstable dual linkage of the proximal and distal carpal rows. A: The midcarpal joint is inherently unstable and tends to assume a lunate-extended posture unless constrained by an intact scaphoid. B: Schematic of the stabilizing effect of the scaphoid. (Adapted from Gilford WW, Bolton RH, Lambrinudi C. The mechanism of the wrist joint, with special reference to fractures of the scaphoid. Guy’s Hosp Report 1943;92:52–59.)
FIGURE 5. Computed tomographic sagittal image of chronic humpback scaphoid deformity.
Horii et al. (50) and Bunker et al. (51) reported a second mechanism for scaphoid fractures that occurs in the so-called punch position of wrist neutral or slight flexion. Based on a group of 18 patients who sustained waist fractures of the scaphoid induced by punching, Horii et al. theorized that axial forces along the second metacarpal with the hand in a full grip would be transmitted to the trapezium and trapezoid, which would in turn impart a palmar-directed shear force to the distal pole of the scaphoid. The fracture morphology in this group of patients was no different than those among patients who sustained hyperextension wrist injuries, implicating the scaphoid’s unique shape and its role as a bridge across the carpus as factors that render it particularly vulnerable to fracture.
FIGURE 6. Dorsal intercalated segment instability deformity with marked lunate extension resulting from chronic scaphoid nonunion.
FIGURE 7. Two most common mechanisms of injury. A: Extreme hyperextension locks the scaphoid into the scaphoid fossa of the distal radius by producing tension in the volar extrinsic ligaments, and the capitate acts as a fulcrum for fracture. B: In the “punch” mechanism, the index metacarpal delivers a palmar-directed shear force across the distal scaphoid via the trapezoid and trapezium. (Redrawn from Horii E, Nakamura R, Watanabe K, et al. Scaphoid fracture as a puncher’s fracture. J Orthop Trauma 1994;8:107–110.)
A historical mechanism for scaphoid fractures was that induced by the crank handle of early automobiles, when the crank suddenly and inexplicably reversed, causing a forced hyperextension and pronation moment on the wrist among those who held on (52) and a direct blow to the anatomic snuffbox among those who released (1). In Leslie and Dickson’s series, this mechanism accounted for 12.5% of their 222 fractures and characteristically caused a transverse waist fracture that required longer immobilization times to heal (1). Verdan theorized that sudden pronation in a hyperextended wrist caused the waist of the scaphoid to experience an acute bending moment across the now taut palmar radiocarpal ligaments, causing a fracture like “a lump of sugar” (52). He described a clinical sign to diagnose an occult scaphoid fracture based on this premise in which the examiner forcibly pronated the patient’s wrist against resistance, a maneuver that produced exquisite pain among those patients with acute fractures and was associated with no discomfort after healing.
P.385

DIAGNOSIS
A high degree of suspicion is necessary to rule out a scaphoid fracture or ligamentous injury after acute hyperextension injuries of the wrist. Often, swelling, pain, and ecchymosis are minimal, and not infrequently, serious injuries are dismissed as minor sprains. Fully 25% of lunate and perilunate fracture dislocations were missed on initial evaluation in one multicenter series (53). If in doubt, the aspiration of blood from an acutely injured wrist is a sign of a serious ligamentous or bony injury, even in light of normal radiographs (54).
Although tenderness in the anatomic snuffbox has been described as a classic physical finding for scaphoid fractures, it is an overly sensitive test that is notoriously inaccurate when used in isolation (55,56). Depending on the quality of radiographs and specialty training of the treating physician, somewhere between 1% and 15% of suspected scaphoid fractures diagnosed in an emergency room by snuffbox tenderness and treated with a 10- to 14-day course of casting are ultimately diagnosed as fractured by follow-up radiographs and/or technetium bone scanning (55,57,58). Although tenderness with axial compression has been suggested as a sign that would increase diagnostic yield (59,60), one study demonstrated 10 out of 25 radiographically evident scaphoid fractures were missed with this test and showed that the test had less than 25% specificity and 41% predictive value among 99 suspected scaphoid fractures (61). In another investigation, tenderness over the scaphoid tubercle was shown to be a sensitive indicator of scaphoid fractures (87%) but had a low specificity (40%), leading these authors to conclude that the tests should be combined to maximize diagnostic ability (55). Parvizi et al. demonstrated that a combination of snuffbox tenderness, scaphoid tubercle tenderness, and pain with axial compression yielded a sensitivity of 100% and a specificity of 70% in a group of 215 suspected acute fractures (56). These authors found any one of the tests in isolation to be inadequate. Waizenegger et al. tested 12 clinical tests in 52 patients with suspected scaphoid fractures and similarly found none of the 12 tests accurate in isolation on initial presentation or at 2 weeks after injury (62).
Because of the inaccuracy of clinical signs, the diagnosis of scaphoid fracture rests on high-quality radiographs. Initial radiographs of the injured wrist should include standard posteroanterior, lateral, oblique, and ulnar deviation posteroanterior grip views (Fig. 8) (10). Because of the similarity of injury mechanisms, an anteroposterior grip film should be obtained to rule out an isolated or concomitant rupture of the scapholunate ligament (63). A “fat stripe,” best seen on an oblique view, is a thin, fat plane that is normally present and parallel to the scaphoid; the stripe is obliterated by a wrist effusion in the case of a scaphoid fracture or serious intercarpal ligament injury (64). If standard views are nondiagnostic, several specialized radiographic views have been investigated. Views that profile the scaphoid in its long axis include a clenched-fist posteroanterior view (65) and an ulnar deviation posteroanterior view in 20 degrees of supination with the beam angled 20 degrees cranially (66). A 45-degree pronated view best profiles fractures of the scaphoid tubercle. A “carpal box” series of magnified and elongated radiographs was demonstrated to be a highly sensitive determinant of laboratory-produced
P.386

scaphoid fractures (67), but a subsequent clinical study of the carpal box series in 71 patients with suspected fracture by the same authors was found to be of limited additional diagnostic value for occult fractures (68). Stereoscopic macroradiography (paired high-resolution magnified imaging) has also been proposed as an adjunct for detection of scaphoid fractures (69), but akin to the carpal box series, macroradiography has been shown to be less than 50% sensitive among occult scaphoid fractures (70).
FIGURE 8. Recommended hand positions for scaphoid fracture series. A: Fisted posteroanterior, lateral, oblique, and semipro-nated, ulnar-deviated view (navicular view). B: Effect of fist on extending the scaphoid to profile the fracture.
FIGURE 9. Value of bone scan in diagnosis of acute scaphoid fracture. A: Ten-day postinjury posteroanterior film fails to demonstrate fracture line. B: Technetium 99m bone scan documents isolated intense uptake in the scaphoid, diagnostic of an acute fracture.
When radiographs do not confirm a fracture, point tenderness in the anatomic scaphoid snuffbox, particularly when associated with one or more additional clinical signs mentioned previously, is presumptive evidence of an occult scaphoid fracture and warrants cast immobilization. A short arm thumb spica cast is recommended for 10 days, at which time, repeat clinical and radiographic examination should be diagnostic (69). The incidence of actual fractures among clinically suspected fractures treated for 10 days in plaster varies from 1.5% to 15% (1,55,57,58,71). Thus, the examiner should be circumspect about ordering additional ancillary imaging studies, which are expensive and in some cases are overly sensitive. Two clinical scenarios warrant the use of ancillary imaging studies: first, if 10-day repeat radiographs are nondiagnostic but persistent tenderness is localized to the scaphoid waist, and second, if a patient’s occupation is such that a 10-day trial of casting would be incapacitating. In these cases, several diagnostic options exist.
Technetium 99m bone scanning has received the most attention in the literature as an adjunct to planar radiographs in the acute follow-up period (57,72,73,74,75,76 and 77). Several authors have demonstrated not only its high sensitivity, which is important to avoid undertreatment of an undisplaced fracture, but also its ability to pick up occult fractures of other carpal bones and/or distal radius (57,78). The bone scan may be positive as early as 48 hours after injury, but accuracy is increased as the interval from injury increases (Fig. 9) (74,78,79). With a sensitivity approaching 100% by 96 hours, technetium bone scanning has a negligible false-negative rate so that immobilization can be confidently discontinued pending a normal study (74,77,78,80,81 and 82). Specificity of a bone scan at 14 days for actual scaphoid fractures among a group of 78 suspected fractures was shown to be 98% by Tiel-van Buul and coauthors and is recommended as their diagnostic procedure of choice for patients with continued snuffbox tenderness and negative radiographs (74).
Ultrasound enjoyed some popularity as a diagnostic modality in the early 1980s (83,84), but subsequent reports refuted its reliability, showing a sensitivity of only 50% (85,86 and 87). Intrasound was heralded by Finkenberg et al. as an accurate and noninvasive technique that required no ionizing radiation in a study of 50 patients with suspected scaphoid fractures (88). These authors demonstrated the technique to both highly sensitive (100%) and specific (95%). In brief, intrasound involves the use of an audible range of vibrating frequencies that is reported to produce discomfort when applied over a fractured carpal bone. A subsequent blinded study by Roolker et al. demonstrated poor diagnostic accuracy with intrasound in a group of 37 patients with suspected occult scaphoid fractures and demonstrated a sensitivity of only 24% (89). Knight and Rothwell replicated the protocol of Finkenberg et al. but showed only 71% sensitivity and 51% specificity in a prospective study of 93 patients in a single hospital and concluded that intrasound vibration had limited usefulness as a diagnostic aid (90).
FIGURE 10. Magnetic resonance imaging of an acute occult scaphoid fracture. A: Plain film is nondiagnostic. B: One-millimeter gradient echo technique demonstrates scaphoid fracture line in mid-waist. C: Fat-suppressed short inversion time inversion recovery images demonstrate edema in the distal pole.
Magnetic resonance imaging (MRI) evolved over the last decade into a highly accurate modality for early diagnosis
P.387

of scaphoid fractures (91,92 and 93). In situations in which immediate diagnosis is critical, Gaebler and coworkers reported that an MRI performed as early as 48 hours postinjury has a sensitivity and specificity approaching 100% and may have the potential to save as much as $7,200 per 100,000 inhabitants by avoiding lost productivity due to unnecessary cast immobilization (92). Other investigators have confirmed a 100% specificity and sensitivity of early MRI, as well as its excellent interobserver agreement (77,93). Most centers recommend a combination of sequences, including T1-weighted spin echo and short inversion time inversion-recovery imaging to maximize diagnostic yield (Fig. 10) (93). Additionally, MRI has the potential to diagnose other bony and soft tissue injuries simultaneously.
Computed tomography has the advantage of a rapid scan time and presents high-resolution fine-cut images of the scaphoid in multiple planes (47,94,95). Computed tomography has all but replaced trispiral tomography in most centers. Sanders described a technique to optimize the plane of sagittal images along the longitudinal axis of the scaphoid by aligning the scout beam within the computed tomography gantry along the longitudinal axis of the scaphoid (Fig. 11) (47). The usefulness of computed tomography for detection of occult scaphoid fractures has not been well established, as computed tomography has been reserved primarily for preoperative analysis of scaphoid nonunion (93). Computed tomography provides unparalleled high-resolution images of bone that can be reformatted in several planes or presented three-dimensionally (Fig. 12) for planning of scaphoid reconstruction. Larsen and colleagues have recommended computed tomography as part of an algorithm to identify carpal fractures should a technetium bone scan be positive in a patient with negative radiographs at 10 to 14 days postinjury (96). Using this protocol, 17 of 31 patients with focal localizing activity on bone scan proved to have distal radius and/or carpal fractures by computed tomography. One study, however, cited a 14% false-negative rate of computed tomography among undisplaced scaphoid fractures that were visualized with plain radiographs (97). With the additional advantage of its ability to identify bone marrow abnormalities, MRI imaging has an edge over computed tomography in the detection of occult scaphoid fractures.
P.388

FIGURE 11. Proper positioning of scout beam and resultant midsagittal scaphoid computed tomography scans. A: The patient lies supine with the shoulder fully abducted and externally rotated over the head. B: Scout beam is aligned parallel to the thumb meta-carpal axis.
FIGURE 12. Computed tomography can be reformatted in several planes to visualize the fracture plane and extent of bone resorption and deformity. A: Dorsal view of scaphoid nonunion. B: Ulnar view of scaphoid nonunion.
Author’s Preferred Method of Diagnosis
Patients who present with a history of an acute hyperextension or axial loading wrist injury undergo a complete radiographic series to rule out carpal fractures, interosseous ligament injury, or perilunate fracture–dislocation (Fig. 13). Unless strongly contraindicated by patient activity level or occupation, patients with point tenderness in the anatomic snuffbox, particularly when accompanied by tubercle and/or axial compression tenderness, are treated in a short arm thumb spica cast to the level of the interphalangeal joint. Ten-day follow-up radiographs should include posteroanterior, lateral, anteroposterior grip, ulnar deviation posteroanterior grip, and 45-degree oblique views. Repeat radiographs are not 100% sensitive for detection of occult fractures, however (93,98). If point tenderness persists in the absence of radiographic findings, a technetium bone scan is the most cost-effective means to rule out a fracture (99). The scan may be performed while maintaining thumb spica immobilization. For those who are unable or unwilling to tolerate a 10-day immobilization period, an MRI scan is diagnostic of a fracture within 48 hours of injury.
FIGURE 13. The author’s preferred algorithm for diagnosis of scaphoid fracture. fx, fracture; MRI, magnetic resonance imaging. (Adapted from Larsen CF, Brondum V, Wienholtz G, et al. An algorithm for acute wrist trauma. A systematic approach to diagnosis. J Hand Surg [Br] 1993; 18:207–212.)
CLASSIFICATION
A scaphoid fracture classification predictive of healing potential would be of the greatest benefit to the clinician. Healing of a scaphoid fracture by closed treatment is most dependent on location of the fracture line and degree of
P.389

displacement. Approximately 20% of scaphoid fractures involve the proximal third (10) and, as noted earlier, are vulnerable to avascular necrosis because of their retrograde blood supply. No stabilizing extrinsic ligaments attach to the proximal pole fracture—it is intimately bound to the lunate by the strong scapholunate interosseous ligament and is subjected to high rotational moments from the ulnar side of the proximal carpal row. Any amount of comminution or displacement increases the potential for fracture site motion and subsequent nonunion (6). As a group, proximal pole scaphoid fractures are responsible for the highest incidence of nonunion, and immobilization times upward of 6 months may be necessary for healing (10,100,101). In contrast, the 10% of scaphoid fractures that occur in the distal third or tubercle heal readily with plaster immobilization, due in part to their excellent vascular supply and in part to a strong cuff of surrounding ligamentous support (Fig. 14) (10,102). The healing potential of waist fractures depends on a number of factors, including comminution, fracture obliquity, vascular supply, and initial displacement.
Weber categorized scaphoid fractures into three groups to better define prognosis for union: nondisplaced fractures, angulated fractures, and displaced fractures (103). The classification system is based on the principles that the displacement of a fracture is proportionately related to the degree of additional soft tissue injury and that the healing potential of the fracture is dependent on its blood supply. Based on laboratory studies, a nondisplaced fracture was the product by a pure bending moment across its waist and was not associated with a disruption of either the interosseous or extrinsic ligaments (104). With increasing energy of injury, dorsal intercarpal ligamentous disruption and/or comminution of the waist induced angulation of the fracture fragments and dorsal tilting of the attached lunate. Complete disruption of extrinsic and/or interosseous ligament attachments to the scaphoid was manifested as displacement of the fracture fragments. Weber and others have shown a high correlation between the appearance of dorsal lunate tilt (dorsal intercalated segment instability) and displacement of the scaphoid fracture fragments (103,105).
P.390

In his analysis of 36 acute scaphoid fractures in a 2-year period, Weber demonstrated a 100% healing rate of 19 nondisplaced fractures; a 67% healing rate and 50% malunion rate of six angulated fractures; and a 55% nonunion rate of eleven displaced fractures (103).
FIGURE 14. Supinated oblique demonstrates a tubercle fracture (arrow) of the scaphoid.
The shape of the scaphoid is so peculiar that any degree of translational or rotational malalignment is readily detected on routine radiographs (Fig. 15). Unless a fracture can be considered anatomically aligned on all radiographic views, the fracture can be considered displaced and unstable and its healing potential via closed means reduced dramatically (54,106). Dorsal angulation of the lunate of 15 degrees or more indicates concomitant injury of the extrinsic carpal ligaments as may have occurred during a combination injury, such as a perilunate fracture dislocation, and operative treatment is mandated.
FIGURE 15. Displaced fracture of the proximal third of the scaphoid.
FIGURE 16. Russe’s classification of scaphoid fracture types. Vertical oblique fractures, although least common, were thought to be more likely to displace during cast immobilization. (Redrawn from Russe O. Fractures of the carpal navicular. J Bone Joint Surg Am 1960;42:759–768.)
Russe (10) and Baumann and Campbell (107) classified fractures depending on the obliquity of the fracture line. These authors recognized that shear forces tend to displace those with a fracture plane aligned vertically, whereas transverse and so-called horizontal oblique fractures were under greater compressive forces and were less likely to displace (Fig. 16). Herbert and Fisher (108) proposed a classification intended to identify those fractures most applicable for operative fixation, either due to inherent instability or delayed/nonunion (Table 1). These authors recommended early operative management of all acute complete fractures of the scaphoid waist or proximal pole because of an alleged propensity to displace or fail to unite in plaster. Despite the assertions of these authors, however, Desai et al. in 1999 were unable to predict fracture union with closed treatment using either the Russe or the Herbert classification systems (109).
Clearly, a large number of variables must be taken into consideration when assessing a scaphoid fracture’s potential to heal by closed means. Although Gellman et al. and others have demonstrated the near universal ability of acute, nondisplaced scaphoid fractures to heal when sufficiently immobilized (1,5,7), a delay in diagnosis, concomitant bony or soft tissue injuries, or the most subtle degree of
P.391

fracture displacement may tip the scale in favor of operative management (1,9,52,54,101,103,106).
TABLE 1. HERBERT AND FISHER’S CLASSIFICATION OF FRACTURES OF THE SCAPHOID (108)
Acute fracture A1 Tubercle fracture
Stable A2 Incomplete waist fracture
Acute fracture B1 Complete waist fracture
Unstable B2 Complete transverse waist fracture
B3 Proximal pole fracture
B4 Transscaphoid perilunate dislocation
Delayed union C Delayed union
Nonunion D1 Fibrous union
D2 Pseudoarthrosis
TREATMENT
Nonoperative Management
Nonoperative management has historically been the mainstay of treatment for scaphoid fractures. Indeed, several studies have documented excellent healing rates with cast immobilization when applied during the acute injury period (1,7,102,103,110,111). Indications for closed treatment of a scaphoid fracture include an isolated, acute, undisplaced fracture of the waist or distal pole. Cast immobilization for subacute injuries is less predictable and extremely dependent on fracture location (19,110).
Pediatric Fractures
Although decidedly rare in the pediatric population, children’s scaphoid fractures do not demonstrate the same propensity to malunion, nonunion, and avascular necrosis as is seen in adults. It has been estimated that scaphoid fractures represent approximately only 0.5% of upper extremity fractures in children (18), and most would agree that they are more difficult to diagnose than to treat (15,22,23). Only eight of 28 fractures were apparent on multiple radiographs taken on the day of injury in one series (22), and the remainder required follow-up radiographs or ancillary imaging at 14 days. MRI has been demonstrated to provide excellent confirmation of fracture status, with a 100% negative predictive rate (112). Virtually all authors report rapid union in all nondisplaced and minimally displaced fractures in an average of 7 to 8 weeks of short-arm cast immobilization (15,17,18,20,21 and 22). Nonunion of untreated scaphoid fractures in children has been reported, and it is recommended that nonunion be treated using a compression screw and bone graft in a similar fashion to adults (Fig. 17) (16).
Adult Fractures
Controversy abounds concerning several variables involved in the closed treatment of adult injuries, however, including the duration of casting, type of cast (long arm vs. short arm), position of immobilization, and inclusion of the thumb interphalangeal joint. Definitive answers to these controversies may never be realized because of the diversity of fracture patterns, location of the fracture lines, individual differences in vascular supply, degree of initial displacement, and associated soft tissue injuries. An assessment of the stability of the fracture may be more critical to fracture healing than many of the casting variables above, as it is likely that stable undisplaced fractures will heal readily within a wide tolerance of wrist position or cast length.
FIGURE 17. Asymptomatic scaphoid nonunion in a 12-year-old patient.
Long Arm vs. Short Arm Cast
Laboratory studies have helped to define the degree to which forearm motion affects movement of the two fragments of a scaphoid fracture. Verdan performed cadaveric studies that demonstrated that rotation of the forearm induced shear forces across the fractured scaphoid through the palmar extrinsic ligaments and recommended extending the cast approximately 3 in. above the elbow to prevent displacement (52). Falkenberg immobilized wrist flexion/extension with an external fixator in cadaveric extremities with simulated scaphoid fractures and demonstrated only minimal fragment motion during forearm rotation (113). Another kinematic study by Kaneshiro et al. used highly accurate stereoradiographic techniques to obtain three-dimensional measurements of fragment rotation and translation and demonstrated up to 4 degrees of rotation and 3.8 degrees of translation of the fracture fragments during forearm rotation (113a). Based on these laboratory studies, it would appear prudent to immobilize the forearm for some period of time, at least until early consolidation of the acute fracture has occurred.
Advocates of both short arm casting and long arm casting, however, cite upward of 95% healing rates with each method of treatment (1,5,9,10 and 11,111). Russe claimed a 97% union rate among 220 acute fractures treated initially with above-elbow casts (10). Langhoff and Anderson reported a 95% union rate among 154 fractures treated within 4 weeks of injury with long arm casting (19). In contrast, Stewart treated 309 military personnel with scaphoid fractures in short arm thumb spica casts and documented just three nonunions (11). Leslie and Dickson obtained union of 95% of 222 patients treated in a short arm cast (1). McLaughlin and Parkes agreed that stable fractures heal equally well in short arm casts, with marked reduction in morbidity (54,101).
“While it is probable that most navicular fractures will unite if well immobilized for a long time, it is equally probable that such triumphs might be reflected in economic
P.392

catastrophes for the young male breadwinners who are most prone to sustain this injury” (54).
Several clinical studies have subsequently compared short- and long arm casting for acute scaphoid fractures. Goldman et al. compared the results of 34 patients treated with a short arm cast with 22 patients treated in an above-elbow cast and found no difference in union rates or time to union (111). Alho and Kankaanpaa randomized 100 consecutive fractures to either a long arm or short arm thumb spica cast and demonstrated a 92% union rate at 12 weeks, with no difference in nonunion rate between groups (9). Neither of these studies excluded displaced scaphoid fractures. Broome et al. retrospectively analyzed healing rates and immobilization times of 31 acute fractures and demonstrated a statistically significant 4-week decrease in healing time in those fractures treated initially in long arm casts (114). Gellman et al. randomized 51 acute nondisplaced scaphoid fractures to either short arm or long arm thumb spica casting for the initial 6 weeks, followed by short arm thumb spica casting until clinical and radiographic union (5). Patients treated with initial long arm cast immobilization showed a significantly shorter time to union (9.5 weeks vs. 12.7 weeks), and the only two nonunions were seen in the short arm cast group. Although it is apparent from these studies that there is no demonstrable difference in union rates between long and short arm cast protocols, the latter two studies support a 6-week period of long arm casting to accelerate union.
Whether or not the thumb is included in the cast, as well as the importance of immobilizing the interphalangeal joint, has been debated as well. At one extreme, an argument to include the entire thumb was advanced by Soto-Hall and Haldeman, who stated that movements of the flexor pollicis longus and abductor pollicis longus were detrimental to scaphoid healing because of their proximity to the fracture line (115). They documented healing in 21 consecutive fractures treated with this protocol (102), and Obletz added an additional series of 45 consecutive scaphoid fractures that united with immobilization in a Soto-Hall position in plaster (116). When treating unstable (vertical oblique) fractures in a cast, Russe used the somewhat extreme concept of a “fist” cast, in which all of the interphalangeal joints were immobilized in flexion (10). Dehne et al. recommended inclusion not only of the thumb interphalangeal joint but also the interphalangeal joints of the radial three digits for additional stability (117). That the thumb interphalangeal joint should be included in the spica cast was questioned by Goldman et al., who cited uneventful healing in 56 of 57 consecutive patients treated with the interphalangeal joint free (111). Neither Cleveland nor Friedenberg believed that the inclusion of the interphalangeal joint or the position of the thumb in the cast made a difference in the healing rate or time to union of scaphoid fractures (118,119). When treating stable fractures, Russe left the thumb completely out of the cast, claiming that thumb motion exerted useful compressive forces across the stable scaphoid fracture site (10). McLaughlin, based on his operative observations of the stability of 19 acute scaphoid fractures during passive wrist motion, concluded that it is the inherent stability of the fracture and not the length or type of the cast that is predictive of healing and favored no immobilization of the thumb at all (101). Clay et al. subsequently randomly allocated 392 acute fractures of the scaphoid to either a short wrist gauntlet without thumb immobilization or a short arm thumb spica cast including the interphalangeal joint and found the incidence of nonunion was independent of the type of cast used (120). For stable fractures treated by cast immobilization, there are insufficient data to conclude that inclusion of all or a portion of the thumb in the cast makes any difference in the rate or time to union of scaphoid fractures (1).
Position of wrist immobilization is a controversial point as well. Weber favored a position of neutral flexion and slight radial deviation (103). He postulated that this position maximally neutralized the force of the radial collateral ligament insertion on the distal fragment, which tended to gap the capitate surface of the scaphoid in a position of wrist ulnar deviation and/or extension. Angulated fractures (without displacement) had a tendency to heal in a malunited, or “humpback,” position that could be diminished by a position of maximum radial deviation and neutral flexion. Soto-Hall and Haldeman favored positioning of the wrist in full radial deviation and 30 degrees of wrist extension and obtained union in all 21 acute scaphoid fractures so treated (102). Goldman et al. documented healing of 55 of 56 patients treated with cast immobilization in wrist extension and slight radial deviation (111). Alho and Kankaanpaa immobilized 100 scaphoid fractures in wrist extension and slight ulnar deviation and reported 92% healing in 12 weeks (9).
In a group of 121 acute scaphoid fractures, Hambidge et al. randomly allocated half the fractures to a 20-degree wrist extension position and half to a 20-degree wrist flexion position (122). Neither group had the thumb immobilized in the cast. The time to union and the 11% incidence of nonunion were not different between the two groups; however, the cohort immobilized in wrist flexion developed a greater loss of wrist extension. Much like the issue of immobilization of the thumb, the importance of wrist position to healing of the fracture may be moot, as its influence is likely overshadowed by the relative stability of the fracture itself.
In an effort to decrease the morbidity of wrist immobilization in athletes, Riester et al. reported union in 10 of 11 middle third scaphoid fractures in collegiate football players treated with a Silastic “playing cast” (123). Only one of three proximal pole fractures healed, and this required many months of immobilization. The authors concluded that a nonrigid playing cast was an acceptable form of treatment for middle third fractures in the high-performance athlete. Rettig et al. reported radiographic healing of 12 of 12 stable scaphoid fractures in athletes treated with a playing cast and concurred that this was a suitable alternative form of treatment for this patient population (124).
P.393

A factor that definitively and adversely affects a scaphoid’s ability to heal with cast immobilization is the duration of time elapsed since injury. Mack et al. reported on a group of 23 subacute fractures, defined as those presenting for treatment between 4 and 24 weeks since injury, and demonstrated a doubling of healing time among undisplaced fractures of the middle third. Among six patients with unstable middle third fractures, two went on to symptomatic malunion or nonunion, and two of three proximal pole fractures failed to heal (110). Langhoff and Andersen reported a near doubling of healing time in fractures diagnosed more than 28 days postinjury and a nonunion rate of 67% among nine proximal pole fractures diagnosed after this interval (19). These studies supported earlier observations by Dickison and Shannon, and Eddeland et al., who reported union rates of only 12% to 23% in fractures more than 8 weeks old (125,126).
Immediate cast immobilization for acute, nondisplaced fractures of the middle and distal thirds of the scaphoid can be predicted to be successful in upward of 95% of patients (1,5,7,10,19,24,100). Absolute contraindications to cast treatment include displacement or angulation in any plane, concomitant perilunate or lunate dislocation (54,127,128,129 and 130), ipsilateral fracture of the distal radius (131), or concomitant scapholunate ligament injury as evidenced by scapholunate diastasis (132) or unilateral lunate dorsiflexion greater than 15 degrees. Relative contraindications to closed management include subacute injuries (greater than 4 weeks since injury), bilateral wrist injury, ipsilateral elbow or shoulder trauma, and multisystem trauma, for which immediate stabilization of the fractured scaphoid would simplify management. Most authors would also consider a small proximal pole fracture an indication for operative stabilization, due to the high incidence of delayed union, nonunion, and avascular necrosis and the considerable morbidity of prolonged immobilization in plaster.
Author’s Preferred Treatment
In low-demand patients with undisplaced fractures of the mid- or distal pole, the author recommends an above-elbow thumb spica cast in wrist extension and neutral forearm rotation for 6 weeks, followed by a short arm wrist gauntlet until healing. There are no absolute guidelines to help distinguish a stable, undisplaced fracture from an unstable, undisplaced fracture (133), hence the author’s recommendation for inclusion of the elbow and thumb for the initial 6 weeks. The author recommends that all subacute fractures, vertical oblique fractures of the scaphoid waist, and all fractures of the proximal pole be treated with early internal fixation. Any degree of displacement, comminution, or carpal malalignment is considered an indication for operative stabilization. Although cast immobilization has long been considered the benchmark for treatment of scaphoid fractures, this view has been challenged by several recent series demonstrating equivalent union rates and improved outcomes after operative intervention and early wrist motion (36,124,134,135 and 136). Young and active patients are increasingly unwilling to consider 3 to 6 months of immobilization when minimally invasive operative management promises the potential for similar outcomes without a cast. The author offers percutaneous fixation (see below) as an alternative form of management, even in nondisplaced fractures that would be expected to heal in a cast.
Follow-Up Imaging Studies
Time to union averages 9 to 12 weeks and is dependent in part on the location of the fracture line, with longer times to healing demonstrated among proximal pole fractures (5,7,133,137). Tubercle and distal pole fractures generally heal within 6 to 8 weeks, middle third fractures in 8 to 12 weeks, and proximal third fractures may require up to 11 months in plaster (6,7 and 8,11,19,137). Casts are generally removed at 4-week intervals and determination of healing made by palpation over the anatomic snuffbox, followed by a scaphoid series of four radiographs. Continued tenderness or demonstration of a persistent radiolucent line is treated with an additional 4-week period of immobilization (5,133). Determination of bony union, however, is notoriously difficult with plain radiographs (47,108,138). Dias et al. found very poor intraobserver and interobserver agreement among experienced examiners concerning healing of scaphoid fractures after 12 weeks of immobilization (139). Herbert and Fisher estimated that up to 50% of scaphoid fractures treated by cast immobilization may ultimately present as an occult nonunion with long-term follow-up (108). Ancillary imaging studies may be required to define union in some cases. Computed tomography, performed with Sander’s protocol in 1-mm cuts along the long axis of the thumb metacarpal, is particularly helpful in determination of bony union (Fig. 18) (47). Partial or incomplete union is not uncommonly seen with computed tomography, and there is evidence to suggest that trabecular bridging on the ulnar side is a more favorable prognostic sign for progression to complete union (140). If computed tomography is not performed to confirm union, radiographs should routinely be repeated 6 weeks and 6 months after release from plaster.
FIGURE 18. Computed tomography to confirm healing. A: Navicular view at 10 weeks demonstrates apparent cortical bridging. B: One-millimeter sagittal cuts demonstrate healing on the radial border. C: Ulnar-sided partial nonunion apparent on contiguous ulnar images.
Operative Management
Techniques for scaphoid fracture fixation continue to evolve. Closed reduction and pin fixation are seldom used because of the difficulties in obtaining either adequate fragment alignment or sufficiently rigid fixation to initiate early range of motion. Smooth Kirschner wire fixation is occasionally useful for nondisplaced fractures in the setting of multiple or open trauma to the hand or upper extremity when rapid fixation is required.
McLaughlin and Parkes (54,130) and others reported their early experience with internal fixation using lag screw fixation for selected acute scaphoid fractures and nonunions through a
P.394

dorsoradial snuffbox approach (Fig. 19). The operative procedure was technically challenging, optimal screw position was not always achieved, and the incidence of nonunion in unstable fractures was not substantially reduced over that obtained by casting alone (34,101,130). Gasser noted the necessity of meticulous dissection to avoid injury to the critical dorsal vascular supply of the scaphoid (34). Because of the size and location of the screw head, hardware removal was generally necessary, and radial sensory nerve injury was not infrequently associated with this operative approach (34,130).
Newer techniques of screw fixation have revolutionized treatment of the fractured scaphoid. Most current implants are headless and designed to be recessed below the articular cartilage. In 1984, Herbert and Fisher introduced the first headless implant with differential leading and trailing thread design intended to provide compression during screw insertion (108). A customized jig facilitated correct placement of the screw and simultaneously maintained external compression on the bony fragments. They reported healing of all 22 acute unstable fractures and recommended against the use of plaster immobilization postoperatively. The authors recognized that the procedure was technically demanding (141), and other authors initially agreed, reporting a high incidence of screw malposition and an unacceptable 42% nonunion rate (Fig. 20) (142). Adams et al. reported an overall healing rate of only 67% among both acute unstable fractures and nonunions treated by Herbert screw fixation but attributed seven of the nine nonunions to technical errors, including inadequate alignment, improper screw length, and inaccurate placement of the guide jig (143). Precise positioning of the implant within the scaphoid was found to be difficult to assess on standard radiographic views (142,144). Chondral penetration by the leading edge of the screw was demonstrated to lead to failure of fixation and articular injury, and fluoroscopy was recommended to decrease the risk of malposition (145).
Further experience with the Herbert screw established its credibility as an excellent device for rapid and secure fracture fixation with low complication rates (36,38,51,146,147 and 148). Bunker et al. demonstrated a 92% union rate among acute fractures and fracture dislocations treated with the Herbert screw in a multicenter prospective study in 1987 (51). These authors underscored the technically demanding nature of the surgery and need for accurate jig placement to avoid malposition.
P.395

FIGURE 19. Exposure and internal fixation of scaphoid fractures through a dorsolateral approach. A: Skin incision. B: Exposure of the dorsal radiocarpal joint capsule after isolating and protecting the superficial radial nerve and radial artery. C: Scaphoid exposure through dorsal radiocarpal capsulotomy. D: Reduction of scaphoid fracture. [Adapted from McLaughlin HL. Fracture of the carpal navicular (scaphoid) bone. J Bone Joint Surg Am 1954;36:765–774.] E: Use of double-guidewire technique for placement of screw and counter-rotation. The second guidewire is removed after placement of the screw.
Freehand placement of a Herbert screw through the dorsoradial approach initially described by McLaughlin (34,101) was reported to achieve 100% union in 10 acute unstable fractures and in 87% of 15 additional cases of delayed or nonunion (38). DeMaagd and Engber reported successful union in 11 of 12 displaced fractures and nonunions of the proximal pole by retrograde placement of a Herbert screw through a limited straight dorsal approach (149). The dorsal approach had the theoretical advantage of minimal disruption of the extrinsic supporting ligaments. Despite a concern for jeopardizing the tenuous scaphoid vascular supply, no increased incidence of avascular necrosis was reported (36,37 and 38,149). Botte et al. demonstrated no substantial interruption of scaphoid vascular supply when a Herbert screw was inserted by either the dorsal or palmar approach in a series of cadaveric wrists, provided some care
P.396

was exercised to avoid the critical ridge vessels (150). The dorsal exposure soon became the preferred method for internal fixation of fractures of the scaphoid proximal pole (39,151), and Rettig and Raskin confirmed the reliability of this approach, with 100% healing rates among 17 acute fractures treated in this fashion (Fig. 21) (36).
FIGURE 20. Malposition of Herbert screw. (Photograph contributed by William Morgan, M.D.)
FIGURE 21. Lateral view of cannulated screw placed through dorsal approach for minimally displaced scaphoid waist fracture.
A cannulated variation on the Herbert screw was introduced by Whipple, designed to be placed under arthroscopic guidance over a single guidewire and incorporating larger trailing threads to increase the internal compressive abilities of the implant (152). Early experience with the device was favorable (153), but there have been limited data comparing the arthroscopic-assisted approach with conventional open management.
Several other devices have been introduced with the intent of combining rigid fixation and interfragmentary compression. Numerous biomechanical studies have compared the compressive ability, bending strength, and pullout properties of the various intramedullary implants (154,155,156,157,158,159 and 160), but no data exist to gauge the requirements for fixation rigidity. Carter et al. determined that both Herbert screws and the AO 3.5-mm cannulated screw were significantly more stable than parallel 0.045-in. Kirschner wires after osteotomy of paired cadaveric scaphoids (158). Other authors have determined that the AO 3.5-mm cannulated screw, Herbert/Whipple screw (Zimmer, Inc., Warsaw, IN), and Acutrak (Acumed, Hillsboro, OR) screws provided significantly increased resistance to cyclic bending loads than the Herbert screw, and these differences were magnified when the devices were compared in a more comminuted model (155,156 and 157). Faran et al. determined that the Acutrak device generated 42% increased compressive force when compared with the Herbert screw but that this effect was very dependent on bone mineral density and increased dramatically in stronger bone (159). Lo et al. questioned the ability of the Herbert screw to maintain compression once the Huene compression jig was removed by noting a 62% drop in compressive force on jig removal in a cadaveric scaphoid osteotomy model (161). The degree to which compression affects union rate, however, cannot be assessed, nor have the actual magnitudes of bending, torsional, and compressive loads across the fractured scaphoid been elucidated.
Trumble et al. (162) compared a cannulated device against the original Herbert screw in two retrospective cohorts of scaphoid nonunion and demonstrated equivalent rates of union but shorter times to healing in the cannulated
P.397

fixation group. These authors attributed the difference to the improved ability of the surgeon to place a cannulated device accurately within the central third of the bone, a factor that was shown to be an independent predictor of more rapid healing. Although not analyzed in a group of acute fractures, accurate placement of the screw within the center of the scaphoid improves biomechanical parameters of fixation and should be similarly expected to positively affect healing.
Devices other than screws have been used for scaphoid fixation, including plate systems and compression staples. Although early results with compression staple fixation for both fresh fractures and nonunions were encouraging, longer follow-up documented a deterioration in clinical outcomes, thought to be secondary to articular damage from the prominent implants (163,164). Huene and Huene advocated the use of the Enders blade plate for scaphoid fractures or nonunions with a small proximal pole, avascularity, or comminution that might limit the holding power of a conventional screw (165). Although a highly stable form of fixation, blade plates carry a high risk of secondary surgery for implant removal because of their tendency for articular impingement (165).
Immediate Internal Fixation
To reduce the morbidity of cast immobilization in active individuals, trials of immediate internal fixation of stable fractures have been performed in selected patients with good results. In 1954, McLaughlin reported on five patients treated with acute internal fixation and immediate range of motion for stable waist fractures; all healed uneventfully (101). Rettig and Kollias reported healing of 11 of 12 scaphoid fractures in athletes treated with immediate internal fixation with Herbert screws through a limited open approach (134). Short arm thumb spica casts were used postoperatively for only 7 to 10 days, and players returned to competitive athletics at an average of 6 weeks postoperatively. Wrist range of motion was symmetric with the opposite side in nine of the 12, and the remaining four lost between 5 and 15 degrees of extension. Rettig et al. treated 18 athletes who were not permitted to compete in a playing cast with immediate internal fixation and showed comparable outcomes to a second group treated with a playing cast, with no difference in healing rates or time to union (124). All patients in the operative group healed primarily without postoperative immobilization and returned to play within 8 weeks. Schroeter et al. reported return to play within 3 to 4 weeks and healing in three of four NCAA basketball players treated with immediate operative fixation and recommended this treatment for selected patients in the high-demand athletic population (166).
Percutaneous Fixation
Although the theoretical advantages of immediate operative stabilization of scaphoid fractures are evident (108,124,134), the disadvantages of an open operative approach for internal fixation of the scaphoid are also well recognized, including the need to expose and divide the palmar radiocarpal ligaments, the potential for interruption of the already tenuous blood supply, as well as the possibility of a hypertrophic and painful scar at the wrist crease (Fig. 22) (148). Garcia-Elias et al. demonstrated that the use of the palmar approach for operative repair of scaphoid nonunions was associated with a statistically significant increase in both scapholunate and lunocapitate angles when compared with the dorsal approach and attributed these markers of carpal instability to the requisite division of the critical palmar extrinsic ligaments (167). Previously, both Dehne et al. (117) and Cosio and Camp (168) demonstrated union rates of 77% and 83%, respectively, among patients with established scaphoid nonunions treated with percutaneous pinning and cast immobilization alone. Based on the premise that some scaphoid fractures failed to unite simply because the fragments were too unstable to be adequately immobilized in plaster, several authors pioneered percutaneous techniques for compression screw osteosynthesis of acute unstable scaphoid fractures (26,54,136,169,170,171 and 172).
FIGURE 22. Scar hypertrophy, contracture, and tenderness are not uncommon after the palmar hockey-stick approach to the scaphoid. In this patient, stiff and swollen digits complicated this 8-week postoperative course.
In 1970, the first trials of percutaneous fixation for acute scaphoid fractures were reported in the German literature by Streli using the 3.5-mm AO screw; however, union rates were no better than contemporary open approaches, and several additional technical difficulties were encountered (169). Schwarz reported his experience with Streli’s technique in 1981 and demonstrated satisfactory results in 21 of 29 patients but continued nonunion in 25% (170). Schwarz recommended the technique in selected complicated scaphoid fractures to reduce prolonged immobilization time but, because of technical challenges and continued high nonunion rates, still preferred cast treatment for most scaphoid fractures.
In 1991, Wozasek and Moser published their experience on 146 acute fractures treated with percutaneous internal
P.398

fixation using a modification of Streli’s technique and cannulated 3.5-mm screws (26). Union occurred at an average of 16 weeks in 130 of the fractures (89%), and patients returned to work at an average of 3 weeks. Stable fractures were not immobilized postoperatively. All screws were removed to avoid styloid or trapezial impingement by the screw head. Five of the 16 nonunions were attributed to technical errors of screw placement, and there were seven additional minor complications. The authors recommended the technique for acute fractures of the scaphoid as an alternative to cast immobilization or open surgery.
A dramatic change in the outcome of percutaneous fixation accompanied the advent and increased use of headless screw fixation. By using the cannulated Herbert-Whipple screw, Inoue and Tamura reported union in ten of ten nondisplaced scaphoid fractures and advocated the procedure for patients who could not afford to have their extremity immobilized, those who required early use of their injured hand, and those with multiple upper extremity injuries (135). Additional experience with cannulated devices by several other authors established this as the preferred technique for percutaneous internal fixation (136,152,171,173,174). Kamineni and Lavy demonstrated that safe access to the scaphoid could be attained through a 1-cm incision centered over the scaphotrapezial joint, with minimal risk to neural or vascular structures (175). Haddad and Goddard reported successful fixation and 100% union in 15 undisplaced scaphoid fractures treated using the Acutrak screw system through a volar approach (136). Operative time averaged only 24 minutes, and there were no perioperative complications. Patients began immediate postoperative wrist motion, time to union averaged 10 weeks, and there was no loss of motion or grip strength when compared with the opposite side at 3 months postoperatively. The authors preferred the Acutrak screw over other cannulated screws because of its headless design and its more favorable biomechanical profile (159).
A theoretical disadvantage of percutaneous fixation via the palmar approach was the possibility that secure fixation of a small proximal pole fracture could not be attained (171). Slade et al. developed a percutaneous dorsal technique that may be used in conjunction with arthroscopic realignment of displaced fracture fragments (176). In a preliminary series of 16 patients, these authors reported 100% union at 12 weeks and cited ease and precision of screw placement as well as more secure fixation of small proximal pole fragments as advantages of their technique.
Author’s Preferred Technique
Operative management of scaphoid fractures has evolved rapidly since the first reports of Herbert screw fixation in 1984 (108). Few young and active patients are content to remain immobilized in a long arm plaster cast when 100% union rates have been reported with a relatively straightforward operative procedure performed through a small incision and followed by immediate range-of-motion exercises (136,171,176). For minimally displaced or nondisplaced fractures of the scaphoid waist in active patients, the author offers percutaneous screw fixation through a volar or dorsal approach as an alternative to cast immobilization. For all acute proximal third fractures, the author advises early internal fixation through a dorsal open or percutaneous approach. For displaced fractures of the distal third or tubercle that are reducible by closed manipulation, percutaneous cannulated fixation through a volar approach is recommended. Acute displaced fractures of the scaphoid waist without comminution can be reduced arthroscopically in some cases and stabilized with a dorsal percutaneous approach. Irreducible fractures of the scaphoid waist are best treated with open reduction and internal fixation with a compression screw through either a dorsal or volar approach. Subacute fractures (more than 4 weeks since injury) are treated with percutaneous or open operative stabilization, as nonoperative treatment has been shown to require an average of 5 months of immobilization (110). Comminuted or angulated fractures and those associated with fracture dislocation of the carpus are treated with a conventional open approach with or without bone grafting and ligament repair (Fig. 23).
Open Technique of Internal Fixation: Palmar Approach
The scaphoid is approached through a 5-cm hockey-stick incision based over the flexor carpi radialis tendon and angled at the scaphoid tubercle. The palmar branch of the radial artery is clamped and ligated. The palmar extrinsic ligaments are exposed and divided sharply and tagged with nonabsorbable sutures for later reapproximation (Fig. 24). The scaphoid fracture fragments are aligned with the use of 0.045-in. Kirschner wires inserted as joysticks. Rarely, severe comminution mandates the use of bone graft, which can be readily obtained by reflecting the pronator quadratus to expose the volar metaphysis of the distal radius. Next, a longitudinal incision is made in the scaphotrapezial capsule and a 3-mm portion of the trapezium cavitated to permit an optimal entry point for the compression screw. Provisional fixation is attained with the use of a single 0.045-in. Kirschner wire placed through the scaphoid tubercle and aimed to penetrate the most dorsal and proximal aspect of the scaphoid. Fluoroscopy is used to confirm central placement of the guidewire on all views (162). If suboptimal, a second guidewire is placed and the first left in place to prevent fragment malrotation during insertion. Care is taken to ensure that chondral penetration by the guidewire has not occurred, and screw length is measured using the depth gauge. The author advises using a screw 5 mm shorter than the measured length from subchondral bone to subchondral bone. The wire is then advanced
P.399

through subchondral bone to prevent inadvertent removal during the drilling process and the cannulated drill advanced by hand or power reaming. The drill is exchanged for the appropriate cannulated screw and final fluoroscopy taken in several planes to confirm placement. If using the Herbert screw, the fracture fragments are held reduced with the Huene compression jig during measurement, drilling, and tapping of the bone, and the tap is exchanged for the appropriate-length screw (Fig. 25). Antirotation wires are removed, the joint irrigated to remove any debris, and the ligaments repaired with nonabsorbable sutures. The author protects stable open repairs with plaster immobilization for 4 to 6 weeks and then initiates range-of-motion exercises in a removable splint. Radiographs are obtained at 6 and 12
P.400

weeks, and computed tomography is used to confirm union as necessary.
FIGURE 23. Author’s preferred algorithm for treatment of acute scaphoid fracture.
FIGURE 24. Conventional open palmar approach to the scaphoid. A: Through a hockey-stick incision, the stout palmar ligaments are divided and tagged to expose the scaphoid waist. Subperiosteal exposure of the distal radius allows concomitant harvesting of cancellous bone graft. B: Close-up view of grafted scaphoid waist fracture. C: Lateral view of internal fixation with cannulated screw.
FIGURE 25. Open placement of Herbert screw. A: Optimal placement of Huene jig and accessory antirotation wire is confirmed with fluoroscopy. B: Herbert screw is advanced under fluoroscopic guidance. C: Central position of the screw is checked in multiple planes.
Open Technique of Cannulated Internal Fixation: Dorsal Approach
Improved fixation is obtained by fixation of proximal pole fractures with a dorsal approach. Sufficient exposure can be gained through a 3- to 4-cm transverse or longitudinal incision based 1 cm distal to Lister’s tubercle. Care is taken to protect any crossing dorsal sensory nerves, and the retinaculum is opened over the third dorsal compartment. The scaphoid is exposed by developing the interval between the wrist extensor tendons and the extensor pollicis longus tendons. The wrist capsule is incised longitudinally and fracture debris and hematoma removed by irrigation. Care is taken to avoid injury to the dorsal ridge vasculature during the approach. The fragments are reduced manually or with the assistance of Kirschner wire joysticks and a provisional wire introduced down the center of the scaphoid while maximally flexing the wrist. The wire is withdrawn distally through the thenar musculature until flush with the proximal articular surface and the reduction and wire placement checked under fluoroscopy. A second wire may be placed to protect against malrotation during screw insertion as necessary. Wire length is measured and the wrist again flexed to advance the wire tip dorsally. The cannulated drill is used under fluoroscopic control. The appropriate length screw is inserted and buried well beneath the articular cartilage for secure purchase in the subchondral bone. Capsular sutures are placed, and the wrist is immobilized until the wound has healed. Early active motion may begin in 7 to 10 days, provided secure fixation is attained.
Arthroscopically Assisted Percutaneous Fixation: Dorsal Approach
Scout guide lines are drawn on the skin in both the frontal and sagittal planes by aligning a 0.045-in. (176) Kirschner wire with the long axis of the scaphoid under fluoroscopic guidance. Manual reduction of the scaphoid fracture fragments is performed under fluoroscopic guidance. A guidewire is prepared by inserting it through the metal cannula of a 14-gauge angiocatheter. With the wrist acutely flexed, the cannula is inserted in the “soft spot” 1 cm distal to Lister’s tubercle and in the location of the 3-4 arthroscopy
P.401

P.402

portal. The correct starting point is confirmed in multiple planes with fluoroscopy. The wire is driven freehand down the center of the scaphoid and out the thenar skin. A portion of the wire is then trimmed to create a second sharp tip dorsally. The wire is withdrawn distally until the proximal wire disappears from view and is buried just below the scaphoid articular cartilage on fluoroscopy. Reduction and correct wire placement are confirmed with fluoroscopy. If the fracture is not anatomically aligned, the wire is withdrawn distally until it clears the proximal fragment, and the fracture fragments reduced manually or with the assistance of additional Kirschner wire joysticks. When satisfactory reduction is attained, the Kirschner wire is advanced back into the proximal pole (Fig. 26).
FIGURE 26. Percutaneous and arthroscopic reduction and fixation of scaphoid. A: Joysticks are placed to facilitate arthroscopic-guided reduction of displaced scaphoid waist fracture. B: Reduction is confirmed under fluoroscopy. C: Antirotation and central wires placed through dorsal percutaneous approach for internal fixation. D: Cannulated screw inserted over guidewire. E: Navicular view of healed fracture at 12 weeks.
FIGURE 27. Percutaneous fixation allows earlier resumption of motion. A: Healed scaphoid fracture at 12 weeks after dorsal percutaneous approach. B: Wrist extension at 12 weeks. C: Wrist flexion at 12 weeks.
The hand is then suspended from finger traps or the arthroscopy tower and 10 to 15 lb of traction applied. Midcarpal arthroscopy is performed to assess the reduction and additional fragment manipulation performed as required. When satisfactory reduction and guidewire placement have been confirmed, the hand is removed from traction and flexed maximally and the wire advanced out of the dorsal skin. A small incision is made alongside the wire using a no. 11 blade, and the scaphoid is reamed using the cannulated drill. Final screw length is measured from the cannulated drill under fluoroscopic control, and the screw is inserted over the guidewire. It is essential to ream at least 1 to 2 mm farther than the intended screw length to avoid distraction of the fracture site that can occur when advancing a screw into unreamed dense bone. The screw should be buried well below the articular cartilage on both ends. The puncture wound is closed with wound closure strips (Steristrips), and plaster immobilization is used for 5 to 7 days to allow swelling to subside. A removable splint is applied, and active wrist motion exercises are begun (Fig. 27).
Percutaneous Fixation: Volar Approach
Traction from finger trap devices or an arthroscopy tower facilitates entrance to the scaphotrapezial joint and an optimal starting point for the internal fixation device (136). The wrist is maximally ulnarly deviated to further open the joint and expose the scaphoid distal pole. A minifluoroscopy unit is very helpful in successful implantation of the device. A 0.045-in. Kirschner wire is placed on the skin to
P.403

identify the optimal angle of wire insertion in both the sagittal and frontal planes, and guide lines are drawn with a skin marker to increase the accuracy of percutaneous wire placement. A 5- to 10-mm incision is made at the point of entry of the guidewire and blunt dissection carried out down to the scaphoid tubercle. A 0.045-in. wire is introduced through the metal cannula of a 14-gauge angiocatheter (serves a dual role as a drill sleeve and wire guide), and the wire/catheter complex is inserted through the incision to a starting point on the scaphoid tubercle (Fig. 28). The wire is driven under fluoroscopic control into the center of the proximal pole. After determination of appropriate screw length, the scaphoid may be predrilled or a self-tapping cannulated screw inserted directly over the Kirschner wire. Final radiographs are obtained and a soft dressing applied. Cast immobilization may be performed at the discretion of the surgeon.
FIGURE 28. Percutaneous scaphoid fixation through a volar approach. With longitudinal traction and ulnar deviation, the guidewire is inserted through a 5-mm incision directly over the scaphoid tubercle. (Redrawn from Haddad FS, Goddard NJ. Acute percutaneous scaphoid fixation. A pilot study. J Bone Joint Surg Br 1998;80:95–99.)
COMPLICATIONS OF SCAPHOID FRACTURE TREATMENT
Complications of cast treatment are well known and include disuse atrophy, stiffness, osteoporosis, pressure sores, and transient pressure neuropathies. Delayed union or nonunion can occur in up to 50% of displaced fractures, and the incidence increases with increased elapsed time since injury.
Operative complications are more common with open approaches to the wrist and include wrist and digital stiffness, infection, hematoma, and hypertrophic scarring. Injuries to dorsal sensory nerves, the distal twigs of the lateral antebrachial cutaneous nerve at the thenar base, or the palmar cutaneous branch of the median nerve can occur and are best avoided by carefully planned operative approaches and meticulous technique. Deep infection, although rare, should be treated with prompt open irrigation and débridement with delayed or secondary wound closure; retention of hardware should be attempted as long as fixation is secure. Accurate implant placement is essential, and great care must be taken to avoid proximal chondral penetration of the implant into the radiocarpal joint space. Fluoroscopy is helpful to ensure accurate fragment reduction and implant placement, and cannulated devices have a shorter learning curve than the Huene compression jig designed for the original Herbert screw.
Grossly improper implant placement should be corrected at once by revision surgery. Loss of fixation is similarly best treated with revision surgery with a compression screw and bone grafting. Failure to adequately reduce interscaphoid angulation predictably leads to malunion, which may be associated with pain and progression of degenerative arthritis (177). Some consideration should be given to early reexploration and bone grafting to restore anatomic alignment.
Delayed healing after operative fixation (more than 16 weeks) is unusual, and loss of fixation or avascular necrosis should be suspected. If internal fixation appears sound on radiographs, bone stimulation with pulsed electromagnetic field application (178) or low-intensity ultrasound (179) may accelerate the healing process, although prospective randomized studies for scaphoid fractures are not available. In established nonunions, the healing rate of pulsed electromagnetic field application and cast immobilization is approximately 70% and is only 50% with proximal pole fractures (178). Low-intensity ultrasound stimulation has been demonstrated both clinically and experimentally to accelerate healing of tibial fractures and distal radius fractures and has been demonstrated to diminish the adverse effects of tobacco on healing (180,181). Mayr et al. demonstrated acceleration of trabecular bridging in a cohort of 30 patients with acute scaphoid fractures treated with cast immobilization and 20 minutes of low-intensity ultrasound application daily when compared with a control group of patients treated by casting alone (182). Union was determined by computed tomography and averaged only 43 days in the experimental group, as opposed to 62 days in the control group. Whether these encouraging results would apply to delayed unions or nonunions cannot be ascertained at this time.
REFERENCES
1. Leslie IJ, Dickson RA. The fractured carpal scaphoid. Natural history and factors influencing outcome. J Bone Joint Surg Br 1981;63:225–230.
2. Osterman AL, Mikulics M. Scaphoid nonunion. Hand Clin 1988;4:437–455.
P.404

3. Praemer A, Furner S, Rice DP. Health care utilization. In: Musculoskeletal conditions in the United States. Park Ridge, IL: American Academy of Orthopaedic Surgeons, 1999:112–113.
4. Hove LM. Epidemiology of scaphoid fractures in Bergen, Norway. Scand J Plast Reconstr Surg Hand Surg 1999;33:423–426.
5. Gellman H, Caputo RJ, Carter V, et al. Comparison of short and long thumb-spica casts for nondisplaced fractures of the carpal scaphoid. J Bone Joint Surg Am 1989;71:354–357.
6. Cooney WP, Dobyns JH, Linscheid RL. Fractures of the scaphoid: a rational approach to management. Clin Orthop 1980;149:90–97.
7. London PS. The broken scaphoid: the case against pessimism. J Bone Joint Surg Br 1961;43:237–244.
8. Mazet R, Hohl M. Radial styloidectomy and styloidectomy plus bone graft in the treatment of old ununited carpal scaphoid fractures. Ann Surg 1960;152:296–302.
9. Alho A, Kankaanpaa U. Management of fractured scaphoid bone: a prospective study of 100 fractures. Acta Orthop Scand 1975;46:737–743.
10. Russe O. Fractures of the carpal navicular. J Bone Joint Surg Am 1960;42:759–768.
11. Stewart MJ. Fractures of the carpal navicular (scaphoid). A report of 436 cases. J Bone Joint Surg Am 1954;36:998–1007.
12. Mack GR, Bosse MJ, Gelberman RH, et al. The natural history of scaphoid non-union. J Bone Joint Surg Am 1984;66:504–509.
13. Ruby LK, Leslie BM. Wrist arthritis associated with scaphoid nonunion. Hand Clin 1987;3:529–539.
14. Mintzer C, Waters PM. Acute open reduction of a displaced scaphoid fracture in a child. J Hand Surg [Am] 1994;19:760–761.
15. Vahvanen V, Westerlund M. Fracture of the carpal scaphoid in children. A clinical and roentgenological study of 108 cases. Acta Orthop Scand 1980;51:909–913.
16. Mintzer C, Waters PM, Simmons BP. Nonunion of the scaphoid in children treated by Herbert screw fixation and bone grafting. A report of five cases. J Bone Joint Surg Br 1995;77:98–100.
17. Wulff RN, Schmidt TL. Carpal fractures in children. J Pediatr Orthop 1998;18:462–465.
18. Christodoulou AG, Colton CL. Scaphoid fractures in children. J Pediatr Orthop 1986;6:37–39.
19. Langhoff O, Andersen JL. Consequences of late immobilization of scaphoid fractures. J Hand Surg [Br] 1988;13:77–79.
20. Greene MH, Hadied AM, LaMont RL. Scaphoid fractures in children. J Hand Surg [Am] 1984;9:536–541.
21. Wilson-Macdonald J. Delayed union of the distal scaphoid in a child. J Hand Surg [Am] 1987;12:520–522.
22. Fechter M, Mayr J, Linhart WE. [Pediatric scaphoid fractures—treatment and prognosis.] Handchir Mikrochir Plast Chir 1998;30:239–242.
23. Larson B, Light TR, Ogden JA. Fracture and ischemic necrosis of the immature scaphoid. J Hand Surg [Am] 1987;12:122–127.
24. Gelberman RH, Wolock BS, Siegel DB. Fractures and non-unions of the carpal scaphoid. J Bone Joint Surg Am 1989;71:1560–1565.
25. Gelberman RH, Gross MS. The vascularity of the wrist: identification of arterial patterns at risk. Clin Orthop 1986;202:40–49.
26. Wozasek GE, Moser KD. Percutaneous screw fixation for fractures of the scaphoid [published erratum appears in J Bone Joint Surg Br 1991;73:524]. J Bone Joint Surg Br 1991;73:138–142.
27. Berger RA. The gross and histologic anatomy of the scapholunate interosseous ligament. J Hand Surg [Am] 1996;21:170–178.
28. Viegas SF, Yamaguchi S, Boyd NL, et al. The dorsal ligaments of the wrist: anatomy, mechanical properties, and function. J Hand Surg [Am] 1999;24:456–468.
29. Moritomo H, Viegas SF, Nakamura K, et al. The scaphotrapezio-trapezoidal joint. Part 1: an anatomic and radiographic study. J Hand Surg [Am] 2000;25:899–910.
30. Taleisnik J, Kelly PJ. The extraosseous and intraosseous blood supply of the scaphoid bone. J Bone Joint Surg Am 1966;48:1125–1137.
31. Gelberman RH, Menon J. The vascularity of the scaphoid bone. J Hand Surg [Am] 1980;5:508–513.
32. Gelberman RH, Panagis JS, Taleisnik J, et al. The arterial anatomy of the human carpus: part I. The extraosseous vascularity. J Hand Surg [Am] 1983;8:367–376.
33. Obletz BE, Halbstein BM. Non-union of fractures of the carpal navicular. J Bone Joint Surg Am 1938;20:424–428.
34. Gasser H. Delayed union and pseudarthrosis of the carpal navicular: treatment by compression-screw osteosynthesis. J Bone Joint Surg Am 1965;47:249–266.
35. Viegas SF. Limited arthrodesis for scaphoid nonunion. J Hand Surg [Am] 1994;19:127–133.
36. Rettig ME, Raskin KB. Retrograde compression screw fixation of acute proximal pole scaphoid fractures. J Hand Surg [Am] 1999;24:1206–1210.
37. Watson HK, Pitts EC, Ashmead D 4th, et al. Dorsal approach to scaphoid nonunion. J Hand Surg [Am] 1993;18:359–365.
38. dos Reis FB, Koeberle G, Leite NM, et al. Internal fixation of scaphoid injuries using the Herbert screw through a dorsal approach. J Hand Surg [Am] 1993;18:792–797.
39. Kozin SH. Internal fixation of scaphoid fractures. Hand Clin 1997;13:573–586.
40. Gilford WW, Bolton RH, Lambrinudi C. The mechanism of the wrist joint with special reference to fractures of the scaphoid. Guy’s Hosp Report 1943;92:52–59.
41. Wolfe SW, Neu CP, Crisco JJ 3rd. In vivo scaphoid, lunate and capitate kinematics in wrist flexion and extension. J Hand Surg [Am] 2000;25:860–869.
42. Linscheid RL, Dobyns JH, Beabout JW, et al. Traumatic instability of the wrist. J Bone Joint Surg Am 1972;54:1262–1267.
43. Smith DK, Cooney WP 3rd, An KN, et al. The effects of simulated unstable scaphoid fractures on carpal motion. J Hand Surg [Am] 1989;14:283–291.
44. Fisk GR. Carpal instability and the fractured scaphoid. Ann R Coll Surg Engl 1970;46:63–76.
45. Smith BS, Cooney WP. Revision of failed bone grafting for nonunion of the scaphoid. Treatment options and results. Clin Orthop 1996;327:98–109.
46. Amadio PC, Berquist TH, Smith DK, et al. Scaphoid malunion. J Hand Surg [Am] 1989;14:679–687.
47. Sanders WE. Evaluation of the humpback scaphoid by computed tomography in the longitudinal axial plane of the scaphoid. J Hand Surg [Am] 1988;3:182–187.
P.405

48. Watson H, Ballet FL. The SLAC wrist: scapholunate advanced collapse pattern of degenerative arthritis. J Hand Surg [Am] 1984;9:358–365.
49. Weber ER, Chao EY. An experimental approach to the mechanism of scaphoid waist fracture. J Hand Surg [Am] 1978;3:142–148.
50. Horii E, Nakamura R, Watanabe K, et al. Scaphoid fracture as a puncher’s fracture. J Orthop Trauma 1994;8:107–110.
51. Bunker TD, McNamee PB, Scott TD. The Herbert screw for scaphoid fractures. A multicentre study. J Bone Joint Surg Br 1987;69:631–634.
52. Verdan C. Fractures of the scaphoid. Surg Clin North Am 1960;40:461–464.
53. Herzberg G, Comtet JJ, Linscheid RL, et al. Perilunate dislocations and fracture-dislocations: a multicenter study. J Hand Surg [Am] 1993;18:768–779.
54. McLaughlin HL, Parkes JC. Fracture of the carpal navicular (scaphoid) bone: gradations in therapy based upon pathology. J Trauma 1969;9:311–319.
55. Freeland P. Scaphoid tubercle tenderness: a better indicator of scaphoid fractures? Arch Emerg Med 1989;6:46–50.
56. Parvizi J, Wayman J, Kelly P, et al. Combining the clinical signs improves diagnosis of scaphoid fractures. A prospective study with follow-up. J Hand Surg [Br] 1998;23:324–327.
57. Murphy D, Eisenhauer M. The utility of a bone scan in the diagnosis of clinical scaphoid fracture [see comments]. J Emerg Med 1994;12:709–712.
58. Mittal RL, Dargan SK. Occult scaphoid fracture: a diagnostic enigma. J Orthop Trauma 1989;3:306–308.
59. Chen SC. The scaphoid compression test. J Hand Surg [Br] 1989;14:323–325.
60. Grover R. Clinical assessment of scaphoid injuries and the detection of fractures. J Hand Surg [Br] 1996;21:341–343.
61. Esberger DA. What value the scaphoid compression test? J Hand Surg [Br] 1994;19:748–749.
62. Waizenegger M, Barton NJ, Davis TR, et al. Clinical signs in scaphoid fractures. J Hand Surg [Br] 1994;19:743–747.
63. Jones WA. Beware the sprained wrist. The incidence and diagnosis of scapholunate instability. J Bone Joint Surg Br 1988;70:293–297.
64. Terry DW Jr., Ramin JE. The navicular fat stripe: a useful roentgen feature for evaluating wrist trauma. Am J Roentgenol Radium Ther Nucl Med 1975;124:25–28.
65. Stecher WR. Roentgenography of the carpal navicular bone. AJR Am J Roentgenol 1937;37:704–705.
66. Ziter FM Jr. A modified view of the carpal navicular. Radiology 1973;108:706–707.
67. Tiel-van Buul MM, van Beek EJ, Dijkstra PF, et al. Radiography of the carpal scaphoid. Experimental evaluation of “the carpal” and first clinical results. Invest Radiol 1992;27:954–959.
68. Roolker L, Tiel-van Buul MM, Bossuyt PP, et al. The value of additional carpal box radiographs in suspected scaphoid fracture. Invest Radiol 1997;32:149–153.
69. Munk B, Frokjaer J, Larsen CF, et al. Diagnosis of scaphoid fractures. A prospective multicenter study of 1,052 patients with 160 fractures. Acta Orthop Scand 1995;66:359–360.
70. Gaebler C, Kukla C, Breitenseher MJ, et al. Limited diagnostic value of macroradiography in suspected scaphoid fractures. Acta Orthop Scand 1998;69:401–403.
71. Jacobsen S, Hassani G, Hansen D, et al. Suspected scaphoid fractures. Can we avoid overkill? Acta Orthop Belg 1995;61:74–78.
72. Waizenegger M, Wastie ML, Barton NJ, et al. Scintigraphy in the evaluation of the clinical scaphoid fracture. J Hand Surg [Br] 1994;19:750–753.
73. Tiel-van Buul MM, van Beek EJ, Broekhuizen AH, et al. Radiography and scintigraphy of suspected scaphoid fracture. A long-term study in 160 patients [see comments]. J Bone Joint Surg Br 1993;75:61–65.
74. Tiel-van Buul MM, van Beek EJ, Borm JJ, et al. The value of radiographs and bone scintigraphy in suspected scaphoid fracture. A statistical analysis. J Hand Surg [Br] 1993;18:403–406.
75. Wilson AW, Kurer MH, Peggington JL, et al. Bone scintigraphy in the management of X-ray-negative potential scaphoid fractures. Arch Emerg Med 1986;3:235–242.
76. Olsen N, Schousen P, Dirksen H, et al. Regional scintimetry in scaphoid fractures. Acta Orthop Scand 1983;54:380–382.
77. Fowler C, Sullivan B, Williams LA, et al. A comparison of bone scintigraphy and MRI in the early diagnosis of the occult scaphoid waist fracture. Skeletal Radiol 1998;27:683–687.
78. Murphy DG, Eisenhauer MA, Powe J, et al. Can a day 4 bone scan accurately determine the presence or absence of scaphoid fracture? Ann Emerg Med 1995;26:434–438.
79. Coupland DB. Determining the presence of scaphoid fracture with a day 4 bone scan. Clin J Sport Med 1996;6:137.
80. Vrettos BC, Adams BK, Knottenbelt JD, et al. Is there a place for radionuclide bone scintigraphy in the management of radiograph-negative scaphoid trauma? S Afr Med J 1996;86:540–542.
81. Tiel-van Buul MM, Roolker W, Broekhuizen AH, et al. The diagnostic management of suspected scaphoid fracture. Injury 1997;28:1–8.
82. Ganel A, Engel J, Oster Z, et al. Bone scanning in the assessment of fractures of the scaphoid. J Hand Surg [Am] 1979;4:540–543.
83. Shenouda NA, England JP. Ultrasound in the diagnosis of scaphoid fractures. J Hand Surg [Br] 1987;12:43–45.
84. Bedford AF, Glasgow MM, Wilson JN. Ultrasonic assessment of fractures and its use in the diagnosis of the suspected scaphoid fracture. Injury 1982;14:180–182.
85. DaCruz DJ, Taylor RH, Savage B, et al. Ultrasound assessment of the suspected scaphoid fracture. Arch Emerg Med 1988;5:97–100.
86. Munk B, Bolvig L, Kroner K, et al. Ultrasound for diagnosis of scaphoid fractures. J Hand Surg [Br] 2000;25:369–371.
87. Christiansen TG, Rude C, Lauridsen KK, et al. Diagnostic value of ultrasound in scaphoid fractures. Injury 1991;22:397–399.
88. Finkenberg JG, Hoffer E, Kelly C, et al. Diagnosis of occult scaphoid fractures by intrasound vibration. J Hand Surg [Am] 1993;18:4–7.
89. Roolker L, Tiel-van Buul MM, Broekhuizen AH. Is intrasound vibration useful in the diagnosis of occult scaphoid fractures? J Hand Surg [Am] 1998;23:229–232.
90. Knight P, Rothwell AG. Intrasound vibration in the early diagnosis of scaphoid fracture. J Hand Surg [Am] 1998;23:233–235.
91. Imaeda T, Nakamura R, Miura T, et al. Magnetic resonance imaging in scaphoid fractures. J Hand Surg [Br] 1992;17:20–27.
P.406

92. Gaebler C, Kukla C, Breitenseher M, et al. Magnetic resonance imaging of occult scaphoid fractures. J Trauma 1996;41:73–76.
93. Breitenseher MJ, Metz VM, Gilula LA, et al. Radiographically occult scaphoid fractures: value of MR imaging in detection. Radiology 1997;203:245–250.
94. Bain GI, Bennett JD, MacDermid JC, et al. Measurement of the scaphoid humpback deformity using longitudinal computed tomography: intra- and interobserver variability using various measurement techniques. J Hand Surg [Am] 1998;23:76–81.
95. Roolker W, Tiel-van Buul MM, Ritt MJ, et al. Experimental evaluation of scaphoid X-series, carpal box radiographs, planar tomography, computed tomography, and magnetic resonance imaging in the diagnosis of scaphoid fracture. J Trauma 1997;42:247–253.
96. Larsen CF, Brondum V, Wienholtz G, et al. An algorithm for acute wrist trauma. A systematic approach to diagnosis. J Hand Surg [Br] 1993;18:207–212.
97. Tiel-van Buul MM, van Beek EJ, Dijkstra PF, et al. Significance of a hot spot on the bone scan after carpal injury: evaluation by computed tomography. Eur J Nucl Med 1993;20:159–164.
98. Tiel-van Buul MM, van Beek EJ, Broekhuizen AH, et al. Diagnosing scaphoid fractures: radiographs cannot be used as the gold standard! Injury 1992;23:77–79.
99. Tiel-van Buul MM, Broekhuizen AH, van Beek EJ, et al. Choosing a strategy for the diagnostic management of suspected scaphoid fracture: a cost-effectiveness analysis. J Nucl Med 1995;36:45–48.
100. Robbins RR, Carter PR. Iliac crest bone grafting and Herbert screw fixation of nonunions of the scaphoid with avascular proximal poles. J Hand Surg [Am] 1995;20:818–831.
101. McLaughlin HL. Fracture of the carpal navicular (scaphoid) bone. J Bone Joint Surg Am 1954;36:765–774.
102. Soto-Hall R, Haldeman KO. The conservative and operative treatment of fractures of the carpal scaphoid (navicular). J Bone Joint Surg Am 1941;23:841–850.
103. Weber ER. Biomechanical implications of scaphoid waist fractures. Clin Orthop Rel Res 1980;149:83–90.
104. Lane LB. The scaphoid shift test [letter]. J Hand Surg [Am] 1994;19:341
105. Smith DK, Gilula LA, Amadio PC. Dorsal lunate tilt (DISI configuration): sign of scaphoid fracture displacement [see comments]. Radiology 1990;176:497–499.
106. Szabo RM, Manske D. Displaced fractures of the scaphoid. Clin Orthop 1988;30–38.
107. Baumann JU, Campbell RD Jr. Significance of architectural types of fractures of the carpal scaphoid and relation to timing of treatment. J Trauma 1962;2:431–438.
108. Herbert TJ, Fisher WE. Management of the fractured scaphoid using a new bone screw. J Bone Joint Surg Br 1984;66:114–123.
109. Desai VV, Davis TR, Barton NJ. The prognostic value and reproducibility of the radiological features of the fractured scaphoid. J Hand Surg [Br] 1999;24:586–590.
110. Mack GR, Wilckens JH, McPherson SA. Subacute scaphoid fractures. A closer look at closed treatment. Am J Sports Med 1998;26:56–58.
111. Goldman S, Lipscomb PR, Taylor WF. Immobilization for acute carpal scaphoid fractures. Surg Gynecol Obstet 1969;129: 281–284.
112. Cook PA, Yu JS, Wiand W, et al. Suspected scaphoid fractures in skeletally immature patients: application of MRI. J Comput Assist Tomogr 1997;21:511–515.
113. Falkenberg P. An experimental study of instability during supination and pronation of the fractured scaphoid. J Hand Surg [Br] 1985;10:211–213.
113a. Kaneshiro SA, Failla JM, Tashman S. Scaphoid fracture displacement with forearm rotation in a short-arm thumb spica cast. J Hand Surg [Am] 1999;24:984–991.
114. Broome A, Cedell CA, Colleen S. High plaster immobilisation for fracture of the carpal scaphoid bone. Acta Chir Scand 2000;128:42–44.
115. Soto-Hall R, Haldeman KO. Treatment of fractures of the carpal scaphoid. J Bone Joint Surg Am 1934;16:822–828.
116. Obletz BE. Fresh fractures of the carpal scaphoid. Surg Gynecol Obstet 1944;78:83–90.
117. Dehne E, Deffer PA, Feighney RE. Pathomechanics of the fracture of the carpal navicular. J Trauma 1964;4:96–114.
118. Friedenberg ZB. Anatomic considerations in the treatment of carpal navicular fractures. Am J Surg 1949;78:379–381.
119. Cleveland M. Fracture of the carpal scaphoid. Surg Gynecol Obstet 1947;84:769–771.
120. Clay NR, Dias JJ, Costigan PS, et al. Need the thumb be immobilised in scaphoid fractures? A randomised prospective trial. J Bone Joint Surg Br 1991;73:828–832.
121. Hofstede DJ, Ritt MJ, Bos KE. Tarsal autografts for reconstruction of the scapholunate interosseous ligament: a biomechanical study. J Hand Surg [Am] 1999;24:968–976.
122. Hambidge JE, Desai VV, Schranz PJ, et al. Acute fractures of the scaphoid. Treatment by cast immobilisation with the wrist in flexion or extension? [see comments]. J Bone Joint Surg Br 1999;81:91–92.
123. Riester JN, Baker BE, Mosher JF, et al. A review of scaphoid fracture healing in competitive athletes. Am J Sports Med 1985;13:159–161.
124. Rettig AC, Weidenbener EJ, Gloyeske R. Alternative management of midthird scaphoid fractures in the athlete. Am J Sports Med 1994;22:711–714.
125. Dickison JC, Shannon JG. Fractures of the carpal scaphoid in the Canadian arm. A review and commentary. Surg Gynecol Obstet 1944;79:225–239.
126. Eddeland A, Eiken O, Hellgren E. Fractures of the scaphoid. Scand J Plast Reconstr Surg 1975;9:234–239.
127. Viegas SF, Bean JW, Schram RA. Transscaphoid fracture/dislocations treated with open reduction and Herbert screw internal fixation. J Hand Surg [Am] 1987;12:992–999.
128. Howard FM, Dell PC. The unreduced carpal dislocation. A method of treatment. Clin Orthop 1986;112–116.
129. Inoue G, Imaeda T. Management of trans-scaphoid perilunate dislocations. Herbert screw fixation, ligamentous repair and early wrist mobilization. Arch Orthop Trauma Surg 1997;116:338–340.
130. Maudsley RH, Chen SC. Screw fixation in the management of the fractured carpal scaphoid. J Bone Joint Surg Br 1972;54:432–441.
P.407

131. Richards RR, Ghose T, McBroom RJ. Ipsilateral fractures of the distal radius and scaphoid treated by Herbert screw and external skeletal fixation. A report of two cases. Clin Orthop 1992;219–221.
132. Vender MI, Watson HK, Black DM, et al. Acute scaphoid fracture with scapholunate gap. J Hand Surg [Am] 1989;14:1004–1007.
133. Barton NJ. Twenty questions about scaphoid fractures. J Hand Surg [Br] 1992;17:289–310.
134. Rettig AC, Kollias SC. Internal fixation of acute stable scaphoid fractures in the athlete. Am J Sports Med 1996;24:182–186.
135. Inoue G, Tamura Y. Closed technique for the Herbert screw insertion in an undisplaced fracture of the scaphoid. J Orthop Surg Tech 1991;6:1–7.
136. Haddad FS, Goddard NJ. Acute percutaneous scaphoid fixation. A pilot study. J Bone Joint Surg Br 1998;80:95–99.
137. Wagner CJ. Fractures of the carpal navicular. J Bone Joint Surg Am 1952;34:774–784.
138. Barton NJ. Apparent and partial non-union of the scaphoid. J Hand Surg [Br] 1996;21:496–500.
139. Dias JJ, Taylor M, Thompson J, et al. Radiographic signs of union of scaphoid fractures. An analysis of inter-observer agreement and reproducibility. J Bone Joint Surg Br 1988;70:299–301.
140. Nicholl JE, Spencer JD, Buckland-Wright JC. Pattern of scaphoid fracture union detected by macroradiography. J Hand Surg [Br] 1995;20:189–193.
141. Herbert TJ, Fisher WE, Leicester AW. The Herbert bone screw: a ten year perspective. J Hand Surg [Br] 1992;17:415–419.
142. Pring DJ, Hartley EB, Williams DJ. Scaphoid osteosynthesis: early experience with the Herbert bone screw. J Hand Surg [Br] 1987;12:46–49.
143. Adams BD, Blair WF, Reagan DS, et al. Technical factors related to Herbert screw fixation. J Hand Surg [Am] 1988;13:893–899.
144. Compson JP, Heatley FW. Imaging the position of a screw within the scaphoid. A clinical, anatomical and radiological study. J Hand Surg [Br] 1993;18:716–724.
145. Tumilty JA, Squire DS. Unrecognized chondral penetration by a Herbert screw in the scaphoid. J Hand Surg 1996;21:66–68.
146. Anderson WJ. Simultaneous fracture of the scaphoid and capitate in a child. J Hand Surg [Am] 1987;12:271–273.
147. Recht J, Evrard H, Guillaume C. [Treatment of fractures of the carpal scaphoid with Herbert’s bone screw. Review of 21 clinical cases.] Acta Orthop Belg 1989;55:183–190.
148. Filan SL, Herbert TJ. Herbert screw fixation of scaphoid fractures [see comments]. J Bone Joint Surg Br 1996;78:519–529.
149. DeMaagd RL, Engber WD. Retrograde Herbert screw fixation for treatment of proximal pole scaphoid nonunions. J Hand Surg [Am] 1989;14:996–1003.
150. Botte MJ, Mortensen WW, Gelberman RH, et al. Internal vascularity of the scaphoid in cadavers after insertion of the Herbert screw [published erratum appears in J Hand Surg [Am] 1989;14:537]. J Hand Surg [Am] 1988;13:216–220.
151. Inoue G, Shionoya K, Kuwahata Y. Ununited proximal pole scaphoid fractures. Treatment with a Herbert screw in 16 cases followed for 0.5–8 years. Acta Orthop Scand 1997;68:124–127.
152. Whipple TL. Stabilization of the fractured scaphoid under arthroscopic control. Orthop Clin North Am 1995;26:749–754.
153. Whipple TL. The role of arthroscopy in the treatment of wrist injuries in the athlete. Clin Sports Med 1998;17:623–634.
154. Kaulesar Sukul DM, Johannes EJ, Marti RK, et al. Biomechanical measurements on scaphoid bone screws in an experimental model. J Biomech 1990;23:1115–1121.
155. Toby EB, Butler TE, McCormack TJ, et al. A comparison of fixation screws for the scaphoid during application of cyclical bending loads. J Bone Joint Surg Am 1997;79:1190–1197.
156. Rankin G, Kuschner SH, Orlando C, et al. A biomechanical evaluation of a cannulated compressive screw for use in fractures of the scaphoid [see comments]. J Hand Surg [Am] 1991;16:1002–1010.
157. Shaw JA. A biomechanical comparison of scaphoid screws. J Hand Surg [Am] 1987;12:347–353.
158. Carter FM II, Zimmerman MC, DiPaola DM, et al. Biomechanical comparison of fixation devices in experimental scaphoid osteotomies. J Hand Surg [Am] 1991;16:907–912.
159. Faran KJ, Ichioka N, Trzeciak MA, et al. Effect of bone quality on the forces generated by compression screws. J Biomech 1999;32:861–864.
160. Newport ML, Williams CD, Bradley WD. Mechanical strength of scaphoid fixation. J Hand Surg [Br] 1996;21:99–102.
161. Lo IK, King GJ, Milne AD, et al. A biomechanical analysis of intrascaphoid compression using the Herbert scaphoid screw system. An in vitro cadaveric study. J Hand Surg [Br] 1998;23:209–213.
162. Trumble TE, Clarke T, Kreder HJ. Non-union of the scaphoid. Treatment with cannulated screws compared with treatment with Herbert screws. J Bone Joint Surg Am 1996;78:1829–1837.
163. Korkala OL, Antti-Poika IU. Late treatment of scaphoid fractures by bone grafting and compression staple osteosynthesis. J Hand Surg [Am] 1989;14:491–495.
164. Korkala OL, Kuokkanen HO, Eerola MS. Compression-staple fixation for fractures, non-unions, and delayed unions of the carpal scaphoid. J Bone Joint Surg Am 1992;74:423–426.
165. Huene DR, Huene DS. Treatment of nonunions of the scaphoid with the Ender compression blade plate system. J Hand Surg [Am] 1991;16:913–922.
166. Schroeter TA, Bassert FH, Strickland JW. Herbert screw fixation of scaphoid fractures in athletes (an earlier return to sport) [abstract]. Orthop Trans 1993;17:439–440.
167. Garcia-Elias M, Vall A, Salo JM, et al. Carpal alignment after different surgical approaches to the scaphoid: a comparative study. J Hand Surg [Am] 1988;13:604–612.
168. Cosio MQ, Camp RA. Percutaneous pinning of symptomatic scaphoid nonunions. J Hand Surg [Am] 1986;11:350–355.
169. Streli R. Perkutane verschraubung des handkahnbeines mit bohrdrahtcompressionsscraube. Zentralbl Chir 1970;95:1060–1078.
170. Schwarz N. Ergebnisse der perkutanen verschraubung des frischen kahnbeinbruches der hand. Unfallheilkunde 1981; 84:302–306.
171. Inoue G, Shionoya K. Herbert screw fixation by limited access for acute fractures of the scaphoid. J Bone Joint Surg Br 1997;79:418–421.
172. Wozasek GE, Moser KD. [Indications for percutaneous screw fixation of scaphoid fractures.] Unfallchirurg 1991;94:342–345.
P.408

173. Ledoux P, Chahidi N, Moermans JP, et al. [Percutaneous Herbert screw osteosynthesis of the scaphoid bone.] Acta Orthop Belg 1995;61:43–47.
174. Taras JS, Sweet S, Shum W, et al. Percutaneous and arthroscopic screw fixation of scaphoid fractures in the athlete. Hand Clin 1999;15:467–473.
175. Kamineni S, Lavy CB. Percutaneous fixation of scaphoid fractures. An anatomical study. J Hand Surg [Br] 1999;24:85–88.
176. Slade JF III, JN Grauer, Mahoney JD. Arthroscopic reduction and percutaneous fixation of scaphoid fractures with a novel dorsal technique. Orthop Clin North Am 2001;32:247–261.
177. Jiranek WA, Ruby LK, Millender LB, et al. Long-term results after Russe bone-grafting: the effect of malunion of the scaphoid. J Bone Joint Surg Am 1992;74:1217–1228.
178. Adams BD, Frykman GK, Taleisnik J. Treatment of scaphoid nonunion with casting and pulsed electromagnetic fields: a study continuation. J Hand Surg [Am] 1992;17:910–914.
179. Kristiansen TK, Ryaby JP, McCabe J, et al. Accelerated healing of distal radial fractures with the use of specific, low-intensity ultrasound. A multicenter, prospective, randomized, double-blind, placebo-controlled study. J Bone Joint Surg Am 1997;79:961–973.
180. Heckman JD, Ryaby JP, McCabe J, et al. Acceleration of tibial fracture-healing by non-invasive, low-intensity pulsed ultrasound. J Bone Joint Surg Am 1994;76:26–34.
181. Cook SD, Ryaby JP, McCabe J, et al. Acceleration of tibia and distal radius fracture healing in patients who smoke. Clin Orthop 1997;198–207.
182. Mayr E, Rudzki MM, Rudzki M, et al. [Does low intensity, pulsed ultrasound speed healing of scaphoid fractures?] Handchir Mikrochir Plast Chir 2000;32:115–122.