The diagnosis of fractures and dislocations of the carpal bones can be difficult for several reasons. The outlines of the eight tightly packed bones are inevitably superimposed in most radiographic views. Even in the anteroposterior view at least one bone overlies another. All views must be interpreted with a thorough understanding of the normal bone contours and the relationships between the bones. Furthermore, the carpal bones normally shift in their relationship to one another during the various arcs of wrist motion.

Because of the difficulty in recognizing fractures in acute injuries, many fractures in this region are not found at initial examination. Articular damage and ligament injuries are even more difficult to evaluate. The latter may permit abnormal rotations and subluxations of the various bones. Special radiographic techniques are helpful. Prognosis is often uncertain because of the peculiarities of the blood supply of these bones, especially of the scaphoid and lunate.


Fracture of the scaphoid bone is the most common fracture of the carpal bones, and diagnosis is frequently delayed. A delay in diagnosis and treatment of this fracture may alter the prognosis for union. A wrist sprain that is sufficiently severe to require radiographic examination initially should be treated as a possible fracture of the scaphoid, and radiographs should be repeated in 2 weeks even though initial radiographs may be negative.


This fracture has been reported in people from 10 to 70 years of age, although it is most common in young adult men. It is caused by a fall on the outstretched palm, resulting in severe hyperextension and slight radial deviation of the wrist. The scaphoid usually fractures in tension. The proximal pole locks in the scaphoid fossa of the radius and the distal pole moves excessively dorsal.  Seventeen percent of patients have other fractures of the carpus and forearm, including transscaphoid perilunar dislocations, fractures of the trapezium, Bennett fractures, fractures of the radial head, dislocations of the lunate, and fractures at the distal end of the radius.


Figure. Mechanism of carpal fractures from falls on outstretched hand with wrist going into marked dorsiflexion.

A, Wrist in marked dorsiflexion. Note that capitate is at 90-degree angle to radius.
B, Scaphoid fractures as result of increased dorsiflexion at midcarpal joint.
C, Dorsal lip of radius strikes capitate, causing it to fracture.
D, Proximal fragment of capitate is rotated 90 degrees.
E, Return of wrist to neutral position. Note that proximal fragment of capitate is now rotated 180 degrees.qq

(From Stein F, Siegel MW: J Bone Joint Surg 51A:391, 1969.)


Anatomy and Blood Supply of Scaphoid Bone


The unique anatomy of the scaphoid predisposes fracture of this carpal bone to delayed union or nonunion and to disability of the wrist. Because it articulates with the distal radius as well as with four of the remaining seven carpal bones, the scaphoid moves with nearly all carpal motions, especially volar flexion. Any alteration of its articular surface through fracture, dislocation, or subluxation or any alteration of its stability by ligamentous rupture can cause severe secondary changes throughout the entire carpus.

The blood supply of the scaphoid is precarious. Only 67% of scaphoid bones have arterial foramina throughout their length, including the distal, middle, and proximal thirds. Of the remaining bones, 13% have blood supply predominantly in the distal third and 20% have most of the arterial foramina in the waist area of the bone with no more than a single foramen near the proximal third. This suggests that one third of scaphoid fractures occurring in the proximal third may be without adequate blood supply, and this seems to be borne out clinically; the prevalence of avascular necrosis can be as high as 35% in fractures at this level. Vessels enter the scaphoid from the radial artery laterovolarly, dorsally, and distally. The laterovolar and dorsal systems share in the blood supply to the proximal two thirds of the scaphoid.

Diagnosis and Treatment

Treatment of scaphoid fractures is determined by displacement and stability of the fracture. Cooney, Dobyns, and Linscheid classified scaphoid fractures as either undisplaced and stable or displaced and unstable. Although this classification remains useful, fractures of the tuberosity, the distal articular surface, and the proximal pole may require special management decisions. For nondisplaced fractures, radiographic diagnosis can be difficult initially. A posteroanterior plain radiograph with the wrist slightly extended in ulnar deviation is helpful. Although repeating radiographs after 2 weeks of immobilization in a cast is a time-honored method for evaluation of a suspected nondisplaced scaphoid fracture, the technetium bone scan and computed axial tomography (in the sagittal plane of the scaphoid) provide diagnostic information sooner.

Nondisplaced, Stable Scaphoid Fractures

Nonoperative treatment usually is successful for acute nondisplaced, stable fractures without other bony or ligamentous injury and for scaphoid fractures in children. The prognosis is better if the fracture is diagnosed early. We use either a forearm cast, or a Munster-type cast, from just below the elbow proximally to the base of the thumbnail and the proximal palmar crease distally (thumb spica) with the wrist in slight radial deviation and in neutral flexion. The thumb is maintained in a functional position, and the fingers are free to move from the metacarpophalangeal joints distally. Using nonoperative casting techniques, the expected rate of union is 90% to 95% within 10 to 12 weeks. Fractures at and distal to the scaphoid waist are expected to heal sooner than those in the proximal pole. During this time, the fracture is observed radiographically for healing. If collapse or angulation of the fractured fragments occurs, surgical treatment usually is required.

If the diagnosis is delayed, or the fracture is in the proximal third, the prognosis is less favorable and an initial long arm thumb spica cast for 6 weeks may be justified. If the diagnosis of an undisplaced fracture of the scaphoid has been delayed for several weeks, treatment should begin with cast immobilization. Surgery should be considered if new healing activity is not evident and if union is not apparent after a trial of cast immobilization for about 20 weeks.

Because of the potential for joint stiffness, muscle atrophy, or the inability to use the hand during and after prolonged immobilization, special nonoperative or operative treatment may be considered in certain patients (e.g., young laborers or athletes). Operative techniques, including percutaneous fixation with cannulated screws, may shorten the time in cast. For some athletes, the use of padded casts during competition may be considered. The advantages and disadvantages of various treatment modifications should be considered in each patient. When making the diagnosis of scaphoid fractures or nonunions, a bipartite scaphoid is considered so rare as to be of little or no clinical significance.

Displaced, Unstable Scaphoid Fractures

A different course of treatment is required for a displaced, unstable fracture in which the fragments are offset more than 1 mm in the anteroposterior or oblique view, or if lunocapitate angulation is greater than 15 degrees, or the scapholunate angulation is greater than 45 degrees in the lateral view (range, 30 to 60 degrees). Other criteria for evaluating displacement include a lateral intrascaphoid angle greater than 45 degrees, an anteroposterior intrascaphoid angle less than 35 degrees (Amadio et al.), and a height-to-length ratio of 0.65 or more (Bain et al.). Because the range of lunocapitate and scapholunate angulation can vary, comparison views of the opposite wrist can be helpful. Reduction can be attempted initially by longitudinal traction and slight radial compression of the carpus. If the reduction attempt is successful, percutaneous fixation with a cannulated screw, or pins and application of a long arm thumb spica cast, may suffice. Otherwise, open reduction and internal fixation may be required.

For a displaced or unstable recent fracture of the scaphoid, the best method of fixation depends on the surgeon’s experience and the equipment available. In some fractures, adequate internal fixation can be obtained with Kirschner wires. The Warner compression staple, the AO cannulated screw, and the Herbert differential pitch bone screw have been used to advantage in displaced and unstable scaphoid fractures. In a comparison study of two groups of patients with scaphoid fractures, Trumble and associates noted 100% union in both groups, one treated with AO cannulated screws and the other with Herbert-Whipple cannulated screws. Cannulated bone screws are useful because the screw can be placed accurately over a guide pin with video fluoroscopic control. The advantages of the Herbert screw, according to Sprague and Howard, are that it (1) reduces the time of external immobilization, (2) provides relatively strong internal fixation, and (3) produces compression at the fracture site. In addition, because the headless screw remains below the bone surface, removal usually is unnecessary. These screws can be used with a bone graft to correct scaphoid angulation. Use of the cannulated design can minimize the disadvantages of the noncannulated screw, which include the use of a jig for insertion and the demanding surgical technique. Contraindications include (1) avascular crumbling of the proximal pole of the scaphoid, (2) extensive trauma or osteoarthritis involving the adjacent carpals or articular surface of the radius, and (3) gross carpal collapse.

Regardless of the fixation device used, careful attention to the details of the procedure, achieving as near an anatomical reduction as possible, and precise placement of the fixation device are of utmost importance.

Malpositioned Nonunion of Scaphoid Fractures (Humpback Deformity)

Established nonunions of scaphoid fractures can be seen in preoperative radiographs to have resorption or comminution, with resulting shortening and angulation, with its convexity dorsal and radial (“humpback” deformity). Preoperative computer-assisted tomography in the sagittal and coronal planes demonstrates this deformity. The deformity includes extension of the proximal pole of the scaphoid, resulting extension of the lunate, and a form of dorsal intercalated instability (DISI) pattern seen on lateral plain radiographs. Interposition bone grafting allows restoration of length and correction of malalignment. Anterior wedge grafting for angulation has been proposed to create a scapholunate angle of more than 60 degrees or an intrascaphoid angle of more than 45 degrees. Fernandez’ modifications emphasized careful preoperative planning, comparison radiographs of the uninjured side, the use of a bone graft fitted to the defect, and Kirschner wire fixation. Tomaino et al. treated persistent lunate extension after interposition grafting of the scaphoid by radiolunate pinning to stabilize the lunate in a neutral position before correcting the scaphoid “humpback” deformity. The cannulated Herbert-Whipple screw was found to be effective fixation. According to Manske, McCarthy, and Strecker, the double-threaded Herbert screw was most effective in nonunions with evidence of avascular necrosis, those involving the proximal third, or those having had previous failed bone grafts. Stark et al. recommended Kirschner wire fixation with an iliac bone graft for all nonunions because judging stability with bone grafting alone was difficult and because the technique was technically easy and added little to the operating time. They achieved union in 97% of 151 old ununited fractures of the scaphoid.


Fractures of the hamate can involve the hamulus, or hook, the body, and various articular surfaces. Fractures of the hook can be treated with casting, open reduction, or excision of the hook. Fractures of the body usually are treated with casting, unless displacement is significant. Articular fractures require open reduction and internal fixation if displacement of the articular surface is 1 mm or more.

A fracture of the hook of the hamate is sometimes difficult to demonstrate. Pain is elicited at the heel of the hand with firm grasp and with pressure against the bony prominence just lateral and slightly distal to the pisiform. A carpal tunnel view may show the fracture (A), but some are better demonstrated by CT (B). When using the latter technique, placing the patient’s hands together in the praying position makes the diagnosis easier because the view of both wrists rules out congenital variation of the hamate, which usually is bilateral. Occasionally the body of the hamate is fractured, but this rarely requires surgery.


A stress fracture may develop in the hook of the hamate with some repetitive activities, such as golf. Initial diagnosis can be difficult. Transient ulnar nerve motor palsy can be caused by an undiagnosed stress fracture of the hook of the hamate. In most instances, unless the diagnosis is delayed, union is likely after immobilization, but excision of the fragment may be necessary for nonunion, persistent pain, or ulnar nerve palsy.


Fractures of the trapezium and trapezoid are rare and may be comminuted when seen in conjunction with radial fracture-dislocations. Displaced trapezial fractures require open reduction. These fractures can be seen radiographically on the carpal tunnel view of the wrist and with tomography or CT scanning.

Palmer classified trapezial fractures into two types:

Type I is a fracture of the base of the ridge, and it may heal when treated by immobilization in plaster;

Type II is an avulsion at the tip of the ridge, and it usually fails to heal when immobilized


Fractures of the trapezium and trapezoid are rare and may be comminuted when seen in conjunction with radial fracture-dislocations. Displaced trapezial fractures require open reduction. These fractures can be seen radiographically on the carpal tunnel view of the wrist and with tomography or CT scanning. Palmer classified trapezial fractures into two types: type I is a fracture of the base of the ridge, and it may heal when treated by immobilization in plaster; type II is an avulsion at the tip of the ridge, and it usually fails to heal when immobilized


Fractures of the lunate can be difficult to detect on plain roentgenography. Tomography and CT scanning may be required to see the fracture. Fractures of the lunate may be nondisplaced, displaced with large fragments, avulsed, especially the dorsal pole, or comminuted. Nondisplaced and nondisplaced comminuted fractures can be treated with cast immobilization. Fractures with more than 1 mm offset and avulsion fractures usually require open reduction. Internal fixation techniques vary depending on the requirements of the individual situation and may include Kirschner wires, small cannulated screws, and suture anchors. Trauma to the lunate may be sufficient to damage the circulation, leading to avascular necrosis of the lunate.

Kienböck disease is a painful disorder of the wrist of unknown cause in which radiographs show avascular necrosis of the carpal lunate. It occurs more frequently between the ages of 15 and 40 years and in the dominant wrist of men engaged in manual labor. In 75% of patients the disorder is preceded by severe trauma, usually with the wrist in severe dorsiflexion. Armistead et al., using CT, demonstrated in some patients occult fractures of the lunate. Untreated, the disease usually results in fragmentation of the lunate, collapse with shortening of the carpus, and secondary arthritic changes throughout the proximal carpal area. Symptoms can develop as early as 18 months before radiographs show evidence of the disease. The use of MRI can be helpful in the diagnosis of early avascular changes in the lunate.

Figure. Common fracture pattern in Kienböck disease is so-called anterior pole type, isolating anterior pole of lunate from remaining portion of bone.

  1.  Distraction of fracture caused by compressive force exerted by capitate diminishes likelihood of fracture healing. This detail usually is not visible on routine roentgenograms because radial styloid process is superimposed on fracture gap. As dorsal portion of lunate collapses further, anterior pole may be extruded volarly.

B. Ratio of height of carpus to length of third metacarpal is reduced in this patient with Kienböck disease. Youm et al. determined that this ratio in normal wrists is 0.54:0.03 and that significantly reduced ratios indicate overall carpal collapse. (From Armistead RB, Linscheid RL, Dobyns JH, Beckenbaugh RD: J Bone Joint Surg 64A:170, 1982)


Lichtman and Degnan’s classification of lunate changes is useful in discussing treatment.

Stage I—Normal architecture with evidence of a linear or compression fracture

Stage II—Normal outline but definite density changes within the lunate

Stage III—Collapse or fragmentation of the lunate and proximal migration of the capitate (carpal height ratio is less than 0.54 ± 0.03)

Stage IV—Generalized degenerative changes within the carpus

The treatment of established Kienböck disease is not standardized. Some have preferred simple casting if the disease is considered to be quite early (stage I or II, before sclerosis, fragmentation, or collapse occurs). Such management includes casting of the wrist for several weeks, if warranted, followed by repeated radiographs in search of occult fracture or avascular changes of the lunate or other disorders that become apparent later, including previously undiagnosed fractures of the carpal scaphoid. This treatment generally has been unacceptable because it requires 4 or more months of immobilization with an uncertain outcome. A study by Mikkelsen and Gelineck, in which 25 wrists were observed for an average of 8 years, concluded that nonoperative treatment of Kienböck disease was ineffective.

Stage I – II.
Some authors recommend ulnar lengthening early in the disease (stage I or II). Hultén has described a condition known as the ulna-minus variant. He found in 78% of patients with Kienböck disease that the ulna was shorter than the radius at their distal articulation. This was true in only 23% of normal wrists. In no patient with Kienböck disease was the ulna longer than the radius at the distal articulation, but 16% of the control group had a so-called ulna-plus variant.

Persson in 1950 reported a series of patients in whom he lengthened the ulna for this disease. These patients were observed for several years by Axelsson and Moberg. They found 16 who had been operated on some 20 years previously, and all but one had been able to continue with manual labor after the operation. Even in one who had pain, the disease process appeared to have been halted. Because of these findings, Armistead et al. have performed the ulnar lengthening operation for Kienböck disease, reporting 20 cases in 1982. Three nonunions required a second plating and bone grafting; 18 of the 20 had pain relief. The technique is detailed below. The ulna should not be lengthened enough to impair ulnar deviation of the wrist; usually most wrist movement can be retained. Strong plate fixation of the distal ulna is recommended.

In addition to ulnar lengthening, the radius can be shortened to provide a level distal radioulnar articular surface for the lunate. Of 12 patients reported by Almquist and Burns, all but one had pain relief, and 10 of 12 showed radiographic revascularization of the lunate. Their indications for radial shortening include negative ulnar variance and lunate compression fracture without fragmentation or flattening (stage II). Earlier studies by Kelven, Axelsson, and Eiken and Niechajev reported similar satisfactory results. Shortening of the radius consists of making a transverse osteotomy about 3 inches (7.6 cm) proximal to the distal articular surface, shortening the radius by 2 mm, and fixing the bone with a compression plate. Edelson, Reis, and Fuchs reported development of Kienböck disease in a patient 16 months after surgery in a lunate that appeared to have reconstituted normally. However, in a series of 35 patients, Schattenkerk, Nollen, and van Hussen reported satisfactory results in two thirds of those treated by both ulnar lengthening and radial shortening. In addition, Weiss et al. reported 30 wrists with Kienböck stages I and II treated with radial shortening of an average of 2.8 mm. At an average 3-year follow-up, 87% had decreased pain, improved motion and grip strength, and no appreciable change in the amount of lunate collapse.

Stage III.
In late cases (stage III) in which the lunate has collapsed but secondary arthritic changes are absent, the ulnar lengthening operation is advocated by Armistead et al. Stark, Zemel, and Ashworth recommended use of a hand-carved silicone rubber spacer in the absence of significant alteration in the shape of the bone, including absence of collapse as measured by the three kinematic indices of McMurtry et al. The carved prosthetic device is substituted for the lunate, which is excised through a dorsal approach. Both Swanson and Lichtman and their associates advocated replacement with a previously molded lunate-shaped silicone block followed by careful repair of the capsule to avoid dislocation of the block. This ligamentous and capsular reconstruction is extremely important and has been emphasized by many authors.

The patient should be warned of the possibility of silicone synovitis and the formation of foreign body cysts. These complications are more likely if the implant is oversized or malpositioned, if carpal instability is present, or if motion or occupational stress of the wrist is excessive. Because of this possibility, some surgeons have abandoned or limited this technique and have suggested intercarpal fusion (scaphoid-capitate, capitate-hamate, or hamate-triquetrum). Simple excision of the lunate, although controversial, has been shown in a study by Kawai et al. to produce satisfactory results with continued relief of pain at an average follow-up of 12 years. In 18 patients the carpus rearranged itself with proximal migration of the capitate, triquetrum, and palmar-flexed scaphoid, but a good range of motion was preserved and degenerative changes were less than anticipated. However, the procedure is not recommended for those who do heavy work. Triscaphe arthrodesis for Kienböck disease has been advocated by Watson. Sennwald and Ufenast evaluated 11 patients treated with scaphocapitate arthrodesis, with an average follow-up of 36 months. They found this procedure effective for pain relief and recommended it for patients with advanced stages of Kienböck disease.

Stage IV.
When secondary arthritic changes have developed throughout the wrist (stage IV), treatment usually is proximal carpal row resection or wrist arthrodesis.