7 Nonoperative Treatment of Scaphoid Fractures
▪ Rationale and Diagnosis
A scaphoid fracture is a common injury, and the usual treatment has been to immobilize the wrist in a plaster of Paris cast. Nonoperative cast treatment of an acute scaphoid fracture achieves union in 88 to 95% of cases.1 , 2 The suggested disadvantages of cast treatment are the period of immobilization, stiffness, decreased grip strength, and delayed return to work.3 , 4 There is an increasing trend toward early fixation of acute scaphoid fractures. This has led to the increasing use of internal fixation, which avoids this period of immobilization and is said to provide a more rapid return to athletic and work activities.5 – 7 It is recognized that fixation of scaphoid fractures can be technically difficult and requires a high level of skill, but few serious complications have been reported.8 – 10 The short- and medium-term results of fracture fixation have not established a significant benefit over traditional cast management.11
The diagnosis of a scaphoid fracture is based on a clinical suspicion reinforced by the history of the injury, a good examination, and appropriate investigations. The injury usually occurs in young men after a forceful fall onto the palm of the hand, which forces the wrist into extension. This can sometimes occur when the hand is violently wrenched backward such as when a starting handle whips back or when the hand is struck on its palm by an object such as a football. The patient usually presents with some pain and swelling on the radial aspect of the wrist. This may not be impressive, and often the patient assumes that the injury is “merely” a wrench. Swelling in the anatomical snuffbox is more common with scaphoid waist fractures. This may not appear for a few days and is best seen by retropulsion of the thumb, which produces a concavity just distal to the radial styloid between the first and third extensor compartment tendons, and then comparing the depth of the concavity on the two sides. Proximal pole fractures and tuberosity fractures are unlikely to cause such swelling. Tenderness to light pressure in the anatomical snuffbox would suggest a bony injury. There is usually a greater than 20% reduction in grip strength that is measured with a Jamar dynamometer. None of these clinical features is sufficiently sensitive or specific in identifying a scaphoid fracture.
In most cases plain radiographs will demonstrate the scaphoid fracture. The size, peculiar anatomy, and oblique orientation of the scaphoid, however, reduce the sensitivity and specificity of a radiographic diagnosis. It is recognized that, unless the x-ray beam lies in the same plane as the fracture, the fracture line may be missed. This has led to a recommendation that multiple radiographic views of the scaphoid be taken. The usual series of four4 views was described by Russe,12 whereas Graziani13 recommended up to 16 different views. Over 90% of scaphoid fractures are detected on a posteroanterior (PA) view or a Ziter14 view of the scaphoid. This view provides an elongated picture of the scaphoid with little overlap of surrounding bones. It is obtained with the wrist in slight flexion and pronation and the hand in maximum ulnar deviation. Very few fractures are picked up solely on the semisupine, semiprone, or lateral views. If uncertain, a computed tomographic (CT) scan or magnetic resonance imaging (MRI) will identify a fracture.
Most fractures involving the waist and distal scaphoid are essentially undisplaced and can be treated nonoperatively regardless of patient age. The ideal candidate for nonoperative treatment is one who can manage to continue working in a below-elbow cast without thumb immobilization for 6 or more weeks or is unwilling to take the risks of immediate surgery.
Avulsion fractures of the tuberosity may not need immobilization in a cast and can be treated in a removable Futuro splint (3M UK PLC, Bracknell, Berkshire, UK), which limits stretching of the distal attachment of the radioscaphoid ligament. Managing fractures in this way imposes a very short period of additional disability but avoids all potential risks of surgery. Large series12 , 15 , 16 have demonstrated the efficacy of conservative treatment of scaphoid fractures with cast immobilization. The aim of restricting wrist movement is to avoid excessive movement of the scaphoid, which may overwhelm the healing efforts of the bone cells at the fracture site. The healing effort reflects the magnitude of the injury, avascularity of the bone fragments, soft tissue laxity, and patient compliance on a back-ground of general health of the individual. Ill health is rarely an issue because this is a fracture of young, fit people.
When a scaphoid fracture is identified, immediate immobilization will facilitate bone healing. Even when a fracture is identified as late as 4 to 6 weeks, it may be treated nonoperatively in a cast, provided that an early CT scan does not demonstrate displacement of the fracture fragments or a clear gap at the fracture site. When the presentation is delayed, patients are treated with careful observation, but the risk of nonunion with cast treatment must be discussed.
There are no absolute contraindications for nonoperative treatment of the broken scaphoid, provided that any significant displacement is reduced. A relative contraindication is a transscaphoid perilunate dislocation. In this situation fixing the scaphoid stabilizes the radial side of the carpus, and attention can then be directed toward the management of the torn lunotriquetral ligament. Because the healing times may be long with cast treatment for proximal pole fractures, early surgical fixation is a reasonable alternative. Significant comminution causing instability and multiple fractures such as a combined scaphoid and radial styloid fracture may need surgical stabilization.
It has been suggested that any displacement or gap is an indication for reduction and fixation. A step, a gap of greater than 1 mm, or angulation at the fracture site of greater than 1515 degrees has been proposed by some as acceptable limits of displacement. In our experience such displacement is uncommon. In 292 fractures reported by Clay et al16 74 had displacement. The union rate for these fractures treated with cast immobilization was 86%. The difference in union rate between fractures considered displaced and those without displacement was small and insignificant. The literature does not provide any information suggesting long-term problems in fractures that have healed with slight displacement, and the contraindication is a theoretical one.17 If the displacement is marked then there is logic in reducing the fracture but this is an uncommon event apart from transscaphoid perilunate dislocation of the carpus.
The least convincing indication for internal fixation is the presence of a bicortical undisplaced fracture.7 The proponents of this treatment suggest that avoiding plaster immobilization allows an earlier return to motion by a few weeks, and the resulting earlier improvement in wrist motion and grip strength is a benefit, which makes the risks of surgery worthwhile. This argument is not compelling and not borne out by comparative studies.
We use a below-elbow cast with the wrist slightly extended and the thumb left free ( Fig. 7.1 ). This position permits the use of the hand while protecting the scaphoid from excessive movement. Most of the scaphoid movement occurs when the wrist moves. Most studies recommending a specific position of the forearm, wrist, or hand are based upon cadaveric investigations. However, in one small prospective study of 51 patients, Gellman et al18 suggested that undisplaced fractures of the scaphoid waist that were initially treated with a long thumbspica cast united as assessed by trabeculae crossing the fracture line on radiographs at 10 weeks, and those that were treated in a short thumbspica cast at 13 weeks.
Also we cannot find evidence that permitting the use of the hand while immobilizing the wrist in the early phases of healing is detrimental to healing. By permitting the use of the hand most patients can return to work provided that the employer permits it and that they can manage. Most individuals, in all but the most demanding tasks, can perform their occupational activities. We advise patients to use their hand as much as possible and to return to have the plaster cast reinforced if it gets soft enough to allow excessive (greater than 10 to 15 degrees) movement of the wrist. All patients are encouraged to resume as much activity as possible, including return to work and leisure activity that does not risk a fall on the outstretched hand or a sudden twist of the wrist.
▪ Potential Pitfalls
To decrease the risk of the cast becoming soft or breaking it is reinforced with a ridge of plaster across the wrist in the front.
Too Much Mobility
In very lax-jointed individuals the thumb may need to be included in the cast to hinder forced pinch and decrease the movement of the scaphoid under pinch load. We also use this method when patient compliance is suspect or the fracture pattern is worrying. We do not immobilize the elbow at any time and also do not attempt to restrict forearm rotation. In this respect our nonoperative management does not compromise the use of the hand and arm as much as when an above-elbow cast18 is used.
Too Early Return to Activity
We carefully counsel our patients on the risk of early return to unprotected loading, such as falling onto the injured hand during a contact sport. Although we are uncertain of the compliance with this advice, we recommend that patients refrain from risky contact sports for around 3 months to allow the fracture healing to consolidate.
In the United Kingdom, the standard treatment of the scaphoid waist fracture consists of immobilization of the wrist in a below-elbow plaster cast with the thumb included up to the interphalangeal joint—the traditional scaphoid plaster. This is said to result in union within 12 weeks in around 95%19 of cases. Despite its wide acceptance the use of the scaphoid plaster is not universal. A simple dorsal slab with the thumb left free was considered adequate by Böhler et al.15 On reviewing 580 of 734 cases accumulated over a 28-year period, a 96.5% union rate had been achieved, and in 35 who had the thumb included in the cast the nonunion rate was even higher. Others have considered inclusion of the thumb to be essential. Soto-Hall and Haldeman20 used a forearm cast with the wrist in extension and radial deviation, and the thumb widely abducted to eliminate the effect of the abductor pollicis brevis, which has part of its origin from the tuberosity of the scaphoid. A 95% union rate was achieved in 4 to 5 months by this method.
Wrist position within the cast is also controversial. Both dorsiflexion and radial deviation of the wrist and dorsiflexion and ulnar deviation have been recommended. Hambidge et al21 found that the wrist position did not influence the union rates (89%) but that at 6 months wrists immobilized in flexion had a greater restriction of extension. Cadaveric studies have suggested that pronation and supination of the forearm caused movement at the fracture site. For this reason several authors have used above-elbow casts. In a CT study of 10 cadaver specimens McAdams et al22 found a rotation of only 0.2 mm at the minimally displaced scaphoid waist fracture site during pronation and supination in a below-elbow cast.
Clay et al16 demonstrated in a large prospective, randomized study that scaphoid fractures treated in a cast that did not immobilize the thumb healed as well as those treated in the conventional scaphoid cast. Although there was no essential difference in the stiffness, tenderness, or functional outcome in the two groups studied, patients treated in the below-elbow cast without including the thumb had better function while in the cast than those treated in the scaphoid cast. Based on this, in our practice only the wrist is immobilized.
The patients are monitored and the cast, hand function, and healing of the fracture are assessed. Union of a fracture can be defined as consisting of bony bridging that restores the normal bony architecture. This is a physiological event that takes a long time, often more than a year, for the bone to form and then remodel so that the bone can bear load normally. The factors that may promote a failure of union are presented in Table 7.1 . The stiffness of the “uniting bone” depends on the time after the fracture and the quality of the healing process. In this process, the bone, at some stage, has healed to an extent that most of the previous ability to bear load and move in a synchronous manner with other carpal bones is restored. This state, when the bone can function with little risk of slowing or arresting healing, is judged as clinical union. Surgery to stabilize the bone alters the state of clinical union. Load transmission is shared between the fixation device and the scaphoid bone. In clinical practice this usually means either having radiographic evidence of callus bridging the fracture site or the absence of adverse radiographic features during the healing period or both ( Table 7.2 ). The only other state is delayed union, which is a diagnosis that can only be made after union has occurred.
Making the diagnosis of a scaphoid fracture on radiographs is difficult because of the shape, size, and orientation of the scaphoid. Trying to establish if the fracture has united is even more difficult. No external callus is expected because most of the surfaces of this bone are articular and are covered with cartilage. Many authors define radiographic union as consisting of bridging trabeculae crossing the fracture line or sclerosis of the fracture line, but these signs are not reliable23 and depend on the angle at which the radiograph is taken. If the x-rays pass through the fracture it will produce a clear gap ( Fig. 7.2 ) at the fracture site on the radiograph. However, if the fracture is at an angle to the x-ray beam the radiograph may show bone bridging even on the first postinjury x-ray.
Union is especially difficult to identify if surgery is undertaken. If good approximation and compression have been achieved there will appear to be “trabeculae crossing the fracture line” immediately after surgery, yet no one would interpret this as union. At 6 to 7 weeks, the same radiographic appearances, which were present immediately after surgery, can now be interpreted as representing union.
If the state of union of the scaphoid fracture cannot be established on radiographic evidence, determining the “time to union” must be a particularly flawed measurement. It depends on the subjective interpretation of unreliable radiographic appearances at an arbitrary interval after surgery.
The implication of measuring time to union is to suggest that the fracture or nonunion will heal in this interval. Measuring time to union requires (1) a clear definition of union, (2) a reliable assessment using current techniques, and (3) a continuous determination of the state of union rather than at the usual intervals after intervention. Although the time to union is documented in several studies this is based on the first visit after intervention when the surgeon has felt able to diagnose union. If a visit had been scheduled a fortnight earlier and the appearances were similar it is conceivable that the time to union could shorten by a fortnight. Alternately if the visit was scheduled a fortnight later or was delayed by the surgeon or patient the time to union would be longer. This is likely to produce very different healing times depending on follow-up schedules. This would add to the unreliability of the assessment of union. The reporting of the time to union should therefore be discouraged.24
Although it is difficult to identify radiographic signs of union of the scaphoid fracture early on it is usually possible to say whether the radiographic images appear satisfactory and that there is no evidence of a failure of union. In the acute scaphoid fracture this is difficult and unreliable because no observable callus is formed. In clinical practice we base the assessment of radiographic union on the absence of adverse features over time, such as a gap at the fracture site or the graft interface, lucency around, or movement of, the implant and displacement of the graft or the fracture.
It is essential to determine that the projection is similar to the image obtained at the previous visit on which the gap was best seen. The shape of the scaphoid on the image and that of any implant such as a screw will provide clues as to whether the image is similar. If the gap disappears over time on similar projections this would suggest that healing is progressing in a satisfactory manner. This gap can also be investigated with a CT scan avoiding metal scatter, or if the implant is made of titanium the gap can be investigated using an MRI scan with a wrist coil. Both the CT scan and the MRI scan provide static images, but they are limited by the type of scanner and the number of slices obtained. Although the presence of a gap suggests a nonunion, the possibility of partial union25 across a sector of the fracture, usually dorsal, should be considered. Such a partial union ( Fig. 7.2 ) may be sufficient for most wrist and hand functions. Although techniques of assessing mobility at the fracture site using real-time ultrasonography26 are possible for conservatively treated fractures of the scaphoid, they are not useful in the presence of implants.
MRI changes in signal intensity or perfusion following immobilization or surgical treatment do not clearly indicate whether there is bony union, but they do provide additional information to support such a conclusion.
The surgeon could, with confidence, state that if no adverse features were identified over 6 months or 1 year it would be safe to assume that the fracture or nonunion had healed ( Table 7.1 and Table 7.2 ).
The interval, either presurgical, postsurgical, or time in weeks after injury or intervention, and the state of individual parameters should be recorded. No attempt should be made to document, analyze, or report time to union because this is a spurious and flawed outcome. If a final outcome is sought this should probably be based on appearances after a minimum interval of 6 months and probably at 1 year. Movement of the fracture and the amount of bony bridging can be determined using a CT scan.25