Clinical Radiology (1993) 48, 398 401

Scaphoid Fracture: A New Method of Assessment D. W. HODGKINSON, D. A. NICHOLSON*, G. STEWART*, M. SHERIDAN* and P. HUGHES*

University Department of Emergency Medicine and *Department of Radiology, Hope Hospital, Salford

The value of colour flow Doppler ultrasound scanning in patients with suspected acute carpal scaphoid fracture has been examined. A prospective study of 78 patients with signs and symptoms of carpal scaphoid injury was undertaken over a period of 6 months. Both wrists of all patients were scanned within 12-72 h of injury. The radial artery and carpal scapboid bone contours are visible using this type of imaging technique. The distance between the medial wall of the radial artery and the outer cortex of the scaphoid waist was measured. The scaphoid index is described and was calculated from this measurement.

Patients were managed without knowledge of the ultrasound imaging result, using plain radiographs and clinical examination as the standard for determining the presence of a scaphoid fracture. All patients were followed to the resolution of symptoms.

All 12 patients who were subsequently considered to have scapboid fractures were identified using this imaging technique within 12-72 h after injury.

We conclude that this imaging may be useful in the assessment of patients with carpal scaphoid fracture. It can be used early, as a supplementary investigation, in patients with suspected carpal scaphoid fracture in whom the initial plain radiographs are normal. Hodgkinson, D.W., Nicholson, D.A., Stewart, G., Sheridan, M. & Hughes, P. (1993). Clinical Radiology 48, 398-401. Scaphoid Fracture: A New Method of Assessment

Accepted for Publication 20 May 1993

The early management of patients with suspected carpal scaphoid fractures, that are not confirmed on the initial plain radiographs, remains unclear. The h~storical view states that the clinical and radiological diagnosis of many scaphoid fractures is difficult or impossible in the first 10 days after injury [1]. This has meant that to this day, X- ray negative suspected carpal scaphoid fractures are managed in a scaphoid plaster for at least 10 days. This is ioappropriate in many cases. In a recent editorial in Injury [2] the author suggested that the time had come to reassess this policy. He mentioned radioisotope scanning as the main development in imaging since the 1970s that could help identify these fractures accurately and early.

We have used an imaging technique, that has not previously been described, to assess patients with sus- pected acute carpal scaphoid fracture. Using colour flow Doppler ultrasound (CFDU) to identify the deep branch of the radial artery, we have measured the displacement of this artery from the outer cortex of the scaphoid waist as it traverses the anatomical snuff box. From this measure- ment, taken in the symptomatic and asymptomatic wrists, we have calculated the scaphoid index which has then been tested for its accuracy in identifying the presence or absence of a carpal scaphoid fracture within 12 72 h after injury.

PATIENTS AND M E T H O D S

Consecutive patients (age > 16 years) presenting to Hope Hospital Accident and Emergency Department between 1 April 1991 and 30 September 1991 with suspected or confirmed carpal scaphoid fracture were included in this prospective study. All patients were initially assessed in the Accident and Emergency Depart-

Correspondence to: D. W. Hodgkinson, University Department of Emergency Medicine, Hope Hospital, Salford M6 8HD.

ment and standard plain radiographs of the scaphoid (PA, PA oblique, lateral, AP oblique pi'ojections) were performed. Patients were managed following standard guidelines with the wrist being immobilized in a plaster cast and arrangements were made for follow-up in the fracture clinic. In addition to this all patients returned to the Accident and Emergency Department between 12 and 72 h after their injury for ultrasound imaging. The radiologist performing the ultrasound was unaware of the results of plain radiographs and all plaster casts were removed prior to imaging and replaced afterwards by the Accident and Emergency Department.

The radial artery traverses the anatomical snuff-box in close proximity to the scaphoid waist separated by only a thin layer of synovium, synovial fluid, wrist joint capsule and radial collateral ligament. The contours of the carpal scaphoid bone and the radial artery are both visible using high resolution ultrasound, but the identification of the artery in particular is considerably improved using colour flow Doppler. An Acuson 128 ultrasound machine with a colour flow facility and utilizing a 5 MHz electronically- steered linear array probe was used to assess all patients. Both the symptomatic and asymptomatic wrists were scanned in the long axis of the scaphoid bone (Figs 1, 2). Optimal images were obtained in this position using high resolution and a single focus to a depth of 2 cm. The scanning plane was located between the tendons of extensor poUicis longus medially and abductor pollicis longus laterally. Abduction of the thumb as shown in Fig. 2 aided orientation along the long axis and improved 'footprint ' access for the probe. A mean value of three measurements was recorded for each wrist. The measure- ment was taken from the anterior cortical reflection of the deepest part of the scaphoid waist to the adjacent medial wall of the radial artery (Figs 3 and 4). Patients were followed up in the fracture clinic, without the knowledge of the ultrasound result, using the usual methods of

/

S C A P H O I D F R A C T U R E 399

Fig. 3 Colour flow Doppler ultrasound image of a normal wrist: note the radial artery (arrow) and the cortex of the scaphoid waist marked with an X.

Fig. 1 -Acuson 128 with 5 MHz linear array probe used to scan an injured wrist.

Fig. 2 Close up of the Acuson probe showing the best position for scanning: note the thumb is in abduction and the wrist is in a neutral position.

clinical assessment and plain radiographs to dictate management . All patients were followed to the resolut ion o f symptoms and informat ion was collected on the length o f t ime spent in plaster and the length o f t ime until discharge f rom the clinic. The clinical data was reviewed together with all the plain radiographs and ul t rasound measurements by the authors.

R E S U L T S

Seventy-eight patients (46 males and 32 females) with a mean age o f 36.8 years (range 16 79 years) were admit ted

Fig. 4 - Colour flow Doppler ultrasound image of an abnormal wrist: note the radial artery (arrow) abnormally separated from the cortex of the scaphoid waist (X).

Table 1 - Scaphoid index calculation

M e a s u r e m e n t s

s y m p t o m a t i c wr i s t = 3.9 m m

- asymptomatic wrist = 1.7 mm

�9 . . . . . Difference (2.2) x 100 = 790/,, Scaphoid lnoex t~l)= Mean (2.8)

to the study. Fo l low-up in format ion was available on all but one pat ient who failed to at tend the fracture clinic or see her General Prac t i t ioner for review. We were unable to trace this patient. There was a large var ia t ion between different patients in the normal scaphoid/radial artery distance (range 1.0-4.9 mm) measured on the asympto- matic wrist. It was therefore decided to analyse the measurements using a ratio o f the symptomat ic and asymptomat ic wrists. We have called this the scaphoid index (SI). This value was calculated from the mean o f the three measurements recorded from each wrist. The difference between the symptomat ic and asymptomat ic wrist was divided by the mean of the measurements for each wrist and this was expressed as a percentage (Table

400 CLINICAL RADIOLOGY

1 0 0 -

e e

8 0 o �9

e e

N 60

X e e "O r

~ .~ 4 0 O

o 20 if)

- 2 0 1 Fractured

0 ~

0 0 � 9

m�9 0

e l e

! ~

4i"ii. 0 0 ~ 0 0

Oe CO0

I Non-fractu red

C

B A

Fig. 5 - The Scaphoid Index (SI) of each patient calculated from colour flow Doppler ultrasound (CFDU) imaging measurements. Open trian- gle represents patients with triquetral fractures. Solid triangle represents patients with distal radius fractures, The two groups (fractured and non- fractured) were identified at the end of the study using all the clinical information and plain radiographs.

Table 2 - Sensitivity and specificity of test at different scaphoid index values

Line Scaphoid i n d e x Sensitivity Specificity (%) (%) (%)

A > 15 I00 56 B > 20 100 62 C > 30 100 74 D > 40 83 80 E > 45 83 82

- The lines labelled al, phabetically are shown in Fig. 5.

1). Ira some patients this was a negative value, when the asymptomatic wrist measurement was greater than the symptomatic wrist measurement.

The standard against which the 'Colour Flow Doppler Ultrasound' method was compared used all the available clinical information and plain radiographs during a final review by the authors. Twelve patients were judged to have a scaphoid fracture and it was possible to identify nine of these on retrospective assessment of the initial plain radiographs taken in the Accident and Emergency Department. In reality only six were identified at initial assessment. Nine patients had fractures of the scaphoid waist and three had fractures of the scaphoid tuberosity. Sixty-six patients were judged not to have a scaphoid fracture. Of the latter group five patients were shown to have other fractures (three undisplaced distal radius fractures and two avulsion fractures of the triquetral bone). Figure 5 is a scatter diagram showing the scaphoid index (SI) of each patient. All the patients with scaphoid fractures, together with the three patients with undis- placed distal radius fractures (shown by blocked trian- gles) had a scaphoid index > 30%. Table 2 shows the sensitivity and specificity of the scaphoid index (using different cut-off values) for predicting the presence of a scaphoid fracture. Table 3 shows the median and range values for the fractured and non-fractured groups. If a

Table 3 - Median and range values of scaphoid index for fractured and non-fractured groups as identified in Fig. 5

Group Seaphoid index

Median Range

Fractured 60.5 30-86 Non-fractured 11.0 - 15-78

Mann Whitney U test, P=0.0001.

scaphoid index of > 30% is used as a cut-off for a positive test then it has a sensitivity of 100% and a specificity of 74%. Patients with a scaphoid index >30% spent significantly longer in plaster (P = 0.0056 Mann Whitney U test; > 30%, median 34 days and range 0 91 days; <30%, median 12 days and range 0-80 days) and a significantly longer�9 time to discharge (P=0.0001 Mann Whitney U test; > 30%, median 47 days and range 10-133 days; < 30%, median 16 days and range 8-109 days). Of the patients with a scaphoid index < 30% that did not have a triquetral fracture, two spent more than 28 days in plaster and more than 48 days to discharge and four others spent more than 70 days to discharge. No patient with a scaphoid fracture required internal fixation of their fracture during the period of follow-up.

There were 61 patients with no evidence of a scaphoid fracture or other fracture and they were assumed to have sustained a soft tissue injury. In this group those with a scaphoid index > 30% did not spend longer in plaster (P=0.52; >30%, median 11 days and range 0-41 days; <30%, median 12 days and range 0-80 days) but did spend longer to discharge (P=0.17; > 30%, median 28 days and range 10-74 days; < 30%, median 16 days and range 8-109 days), the latter value was not significant.

CF D U identified all 12 patients who were judged to have a carpal scaphoid fracture and particularly the six patients in whom this diagnosis was not made at initial assessment (three of whom had normal initial plain radiographs).

DISCUSSION

The diagnosis of scaphoid fracture by plain radio- graphs or isotope bone scanning remains a matter for debate [3]. In this study 25% [3] of patients with a scaphoid fracture had a normal standard series of plain radiographs of the scaphoid at the time of presentation. In a further 25% the initial plain film abnormality was only identified at review. This figure is similar.to that found by Mazet and Hohl [4]. Leslie and Dickson [5] suggested that only 2% of scaphoid fractures cannot be seen on the initial standard radiographs as long as they are of good enough quality to reveal the trabecula pattern of the bone. Some authors have described the scaphoid fat stripe as being a useful pointer to the presence of a scaphoid fracture [6,7]. Its false negative and false positive rates have been shown to be too high for it to be used as a discriminating investigation [8]. Tiel-Van Buul et al. [9] investigated the radio-isotope scanning of suspected scaphoid fractures and concluded that if a positive bone scan correlated with a scaphoid fracture then 65% of all scaphoid fractures were missed by plain radiographs. This figure probably overestimates the problem since false positive bone scans are not uncommon in the young

SCAPHOID FRACTURE 401

and elderly populat ions; however, plain radiographs will miss a p ropor t ion o f patients with a scaphoid fracture. We have described an ul t rasound imaging technique that has previously not been used to investigate this type o f injury and assessed its ability to recognize patients with scaphoid fractures within 12-72 h after injury. This ul t rasound imaging technique is based on the same principle as the scaphoid fat stripe in that y o u would expect swelling or a haemarthrosis t 9 develop a round the scaphoid bone within 12-72 h after fracture. We have shown that this investigation can identify patients who have a scaphoid fracture if a scaphoid index o f > 30% represents a positive result (sensitivity 100%, specificity 74%). The complicat ion rate following X-ray negative scaphoid fractures is very low and Duncan and Thurs ton [10] questioned whether these patients need to be treated in a plaster cast at all. Soft tissue injuries, however, should not be ignored as a certain number o f X-ray negative suspected scaphoid fractures represent severe l igament disruption which in the case o f dorsal intercalated segmented instability may lead to prolonged disability. There is a suggestion that the scaphoid index may identify the more significant soft tissue injuries. The patients with no evidence o f a carpal scaphoid fracture (61) that had an SI > 30% did not spend any longer in plaster but did spend longer to discharge (i.e. longer to resolution o f symptoms) but this was not significant. A larger s tudy is required to clarify this point.

In conclusion we agree with Staniforth 's review [2] o f this subject which states that an early accurate diagnosis is essential and that this should be made as soon as possible by an appropr ia te expert who can examine the patient, assess good quality plain radiographs and where necess- ary order special investigations. We believe that the scaphoid index may be used in the early assessment o f

patients with suspected carpal scaphoid fractures whose plain radiographs are normal, to predict accurately the presence or absence o f a scaphoid fracture,.It can identify scaphoid~fractures and may be able to identify the more significant soft tissue injuries. The use o f this measure- ment may help to reduce the number o f patients with soft tissue injuries who a r e inappropriately immobilized in a scaphoid plaster.

R E F E R E N C E S

1 Russe O. Fracture of the carpal navicular - diagnosis, non-operative treatment and operative treatment. Journal of Bone and Joint Surgery 1960;42A:759.

2 Staniforth P. Scaphoid fractures and wrist pain- time for new thinking. Injury 1991;22:435-436.

3 Dias JJ. Suspected scaphoid fractures. The value of radiographs. Journal of Bone and Joint Surgery 1990;72B:98-10 I.

4 Mazet R, Hohl M. Fracture of the carpal navicular. Analysis of ninety-one cases and review of the literature. Journal of Bone and Joint Surgery 1963;45A:82 112.

5 Leslie l J, Dickson RA. The fractured carpal scaphoid; natural history and factors influencing outcome. Journal of Bone and Joint Surgery 1981;63B:225 230.

6 Cetti R, Christensen SE. The diagnostic value of displacement of the fat stripe in fractures of the scaphoid bone. Hand 1982; 14:75.

7 Curtis DJ. Injuries of the wrist: an approach to diagnosis. Radiologic Clinics of North America 1981;19:625-644.

8 Dias J J, Findlay PB, Brenkel I J, Gregg PJ. Radiographic assessment of soft tissue signs in clinically suspected seaphoid fractures. Journal of Orthopaedic Trauma 1987;1(3):205-208.

9 Tiel-Van Buul MMC, Van Beck E JR, Broekhuizen AH, Nooitge- dacht EA, Davids PHP, Bakker AJ. Diagnosing scaphoid fractures: radiographs cannot be used as a gold standard. Injury 1992;23(2):77-79.

10 Duncan DS, Thurston AJ. Clinical fracture ofthe carpal scaphoid - an illusionary diagnosis. Journal of Hand Surgery 1985; 10B(3):375- 376.