Various classifications of scaphoid fractures have

The anatomy of acute scaphoid fractures A THREE-DIMENSIONAL ANALYSIS OF PATTERNS J. P. Compson From the United Medical and Dental Schools of Guy s and St Thomas Hospitals, London, England Various classifications of scaphoid fractures have been based on plain radiography, but there are difficulties in defining the actual fracture line without an appreciation of the three-dimensional anatomy. Radiological fracture lines were therefore mapped on transparent methylmethacrylate models of the bone. An analysis of 91 acute fractures showed that 11 were apparently incomplete. The other 80 showed three basic anatomical patterns: transverse through the waist, oblique in the plane of the dorsal sulcus, or of the proximal pole. There was some variation and comminution in these patterns, but no distal fractures of the body were seen. The interpretation of different radiological projections is discussed. The findings have implications for the management and the assessment of outcome. J Bone Joint Surg [Br] 1998;80-B:218-24. Received 27 May 1996; Accepted after revision 30 October 1997 Fractures of the scaphoid may involve the body, the tubercle 1-3 or be avulsions of bone fragments. 4,5 Despite the many reports on the clinically important fractures through the body, these injuries are poorly understood because the radiological fracture lines are difficult to relate to the threedimensional anatomy. This has led to problems in appreciating the natural history, the mechanics of collapse and the optimum surgical fixation of the various patterns of fracture. The natural history of fractures of the body must depend on the type of injury. Classifications have usually been based on a limited number of radiological views, 6-17 but the complex shape and orientation of the normal bone mean that these radiographs give distorted and confusing images. Plain radiographs, particularly single views, are very unreliable in determining the true level and alignment of a fracture. Confusion is even greater when there is displacement of an acute injury, or bone loss and collapse after J. P. Compson, FRCS Orth, Consultant in Trauma and Orthopaedics King s College Hospital, Denmark Hill, London SE5 9RS, UK. 1998 British Editorial Society of Bone and Joint Surgery 0301-620X/98/26926 $2.00 nonunion. Most classifications which are based on a few views cannot actually group the fractures into true types and therefore are not able to predict the natural history. This is also true of some clinically based classifications 18,19 which, although useful, do not extend knowledge. Detailed understanding of the alignment of the fractures to the long axis of the bone and their relationship to ligament attachments would help to establish potential stability. This is important because nonunion is closely related to displacement and collapse as well as to loss of bone, 20-22 and knowledge of which types of fracture affect stability would be of value in predicting outcome. Understanding of fracture alignment is also important for effective fixation both of acute injuries and after nonunion. After an analysis of the anatomy of the scaphoid (Fig. 1), including study of the projection of anatomical landmarks on standard scaphoid radiographs, 23,24 a method of delineating the line of fracture was devised, using three-dimensional methylmethacrylate models. Material and Methods After the analysis of 50 dry adult Caucasian scaphoids for shape, anatomical features and variability, ten clear methylmethacrylate models were made of ten bones which covered the full range of variation. The anatomical landmarks seen on standard radiological views were then matched to the appropriate models; their transparency meant that they could be overlaid on radiographs to allow easy visualisation of the bony landmarks. The transparency of the models mimicked the radiological views in showing both volar and dorsal landmarks. A series of radiographs of scaphoid fractures was then analysed, using an appropriately selected model for each case. It is impossible to determine from a single film whether an apparent fracture line is on the front or back of the image, and therefore the fracture lines on each view were marked on the model only at points where they broke the edge (Fig. 2). A single good radiograph provided two definite cortical points, when both ends of the line were visible and not obscured by overlap from other carpal bones or other parts of the scaphoid. For example, in a neutral posteroanterior (PA) view the tubercle overhangs the waist. 218 THE JOURNAL OF BONE AND JOINT SURGERY

THE ANATOMY OF ACUTE SCAPHOID FRACTURES 219 Fig. 2 Method of marking a fracture line on a selected model from a 45 PA oblique view. Fig. 1 Fig. 3 Osteology of a typical scaphoid in four views. The anatomical waist is shown by a bold line (DA, dorsal apex of the ridge; LA, lateral apex of the ridge; R-S, radioscaphoid joint; S-C, scaphocapitate joint; S-L, scapholunate joint; ST, scaphotrapezial joint; S-T-T, scaphotrapeziotrapezoid joint; TUB, tubercle). Lateral views of three scaphoid models, showing the lines of the three fracture patterns: (left) waist; (middle) dorsal sulcus type 3; and (right) proximal pole. Table I. Recommended radiological views for the detection and classification of scaphoid fractures Neutral PA Ulnar-deviated PA Ulnar-deviated PA with 20 tube angulation to the elbow* Misleading, but the usual view taken of a painful wrist Helpful, especially for fractures of the proximal pole Specific view for fractures of the waist: the beam at right angles to the long axis Patterns oblique to the beam are poorly visualised 45 oblique PA (semipronated) Best for sulcal fractures, but also shows waist and tubercle fractures. Best for displacement of waist fractures and some sulcal and proximal-pole fractures 45 oblique AP (semisupinated) Specific view for fractures of the proximal pole. Also shows dorsal avulsions. Best view for assessing flexion in the humpback scaphoid Lateral Poor for fracture detection unless penetration is good. Shows alignment, and its main use therefore is to demonstrate carpal collapse * Bridgeman. 32 Other views showing the long axis are equally acceptable 33,34 Study of a series of views of one fracture allowed it to be plotted circumferentially around the scaphoid model. This process transferred the fracture line from two-dimensional radiographs to the three-dimensional model (Fig. 3). A total of 91 acute scaphoid fractures in adults was analysed, excluding small avulsion fractures around the scapholunate joint and fractures of the tubercle. Six views of each wrist were available for 44 fractures studied prospectively over a two-year period (Table I). Three of these 44 had required internal fixation for significant and increasing fracture displacement. The other 47 scaphoid fractures were studied retrospectively but all had adequate radiographs with at least four views at the time of injury. All these patients had undergone surgery and could be traced from the theatre records of four hospitals. For each of the 91 fractures the model closest to the shape of the scaphoid was selected and the outline of the fracture marked as described above. The fracture patterns seen in the whole series were then analysed to derive a classification. VOL. 80-B, NO. 2, MARCH 1998

220 J. P. COMPSON Table II. Details of the 91 acute fractures of the scaphoid which were studied Prospective* Retrospective (n = 44) surgical (n = 47) Unclassifiable 11 None Waist 11 13 Sulcal Type 1 7 6 2 4 } 15 5 } 14 3 4 3 Proximal pole 7 20 * one sulcal and two waist fractures required internal fixation because of increasing displacement Results Fracture patterns. In 11 of the 47 retrospective cases, it proved impossible to define fully the three-dimensional pattern. In eight of these the fracture line appeared to break only a single cortex and could not be traced around the bone. In the other three the fracture line was doubtful despite a positive isotope scan and a clear clinical diagnosis. All 11 fractures united and it was uncertain whether the imaging had been inadequate, or the fracture was only partial as suggested by McLaughlin and Parkes 18 and Herbert and Fisher. 19 In the 80 scaphoids for which a fracture line could be accurately defined on the appropriate methacrylate model, there appeared to be three main fracture patterns (Fig. 3; Table II) involving: 1) the surgical waist ; 2) the dorsal sulcus; or 3) the proximal pole. Waist (24; Fig. 4). Only one fracture was in the exact plane of the true anatomical waist, but the other 23 varied only slightly, by passing proximal rather than distal to the lateral apex of the dorsal ridge. Apart from this small variation, all Fig. 4a Fig. 4b Fig. 4c Waist fracture: a) the anatomy in four views (LA = lateral apex of the ridge); b) lateral view with the fracture line partly obscured by the capitate; c) ulnar-deviated PA view with 20 angulation of the beam; d) 45 oblique PA view; and e) standard PA view of the same fracture. Fig. 4d Fig. 4e THE JOURNAL OF BONE AND JOINT SURGERY

THE ANATOMY OF ACUTE SCAPHOID FRACTURES 221 Fig. 5b Fig. 5a Dorsal sulcal fracture: a) the anatomy; b) the three variations on a dorsal view (+, lateral apex of ridge); and c) 45 oblique PA view showing a type-3 fracture. Fig. 5c the fractures followed the anatomical waist at right angles to the long axis of the bone. In a few cases, there was some comminution at the lateral apex and in two cases part of the lateral apex was displaced. Dorsal sulcus (29; Fig. 5). Fractures in the line of the dorsal sulcus are at 45 to the surgical waist and therefore to the long axis of the bone. The fractures did not follow the sulcus to the tip of the tuberosity, but crossed the palmar face to the scaphocapitate facet. At the lateral apex of the dorsal ridge there were three variations: the fracture line could pass proximally, distally or on both sides of the apex (Fig. 5b). The butterfly fragment in the third subgroup showed displacement and comminution. Proximal pole (27; Fig. 6). These fractures were proximal to the dorsal ridge and originated at the dorsal apex of this ridge, crossing the radioscaphoid joint and then the scaphocapitate joint near its proximal end. There was considerable variation in the relationship of the fracture to the tip of the proximal pole, which seemed to depend on the proximal extent of the facet with the capitate (Fig. 6b). When this facet nearly reached the tip, the fracture line tended to be more proximal. In some scaphoids with a well-developed proximal half, the scaphocapitate facet is distally placed and a fracture of the proximal pole will seem to be more vertical, with a larger proximal fragment on a PA view. The scaphocapitate facet is one of the most variable features of the scaphoid anatomy, and therefore proximal pole fragments also varied in size, although the anatomical basis remained constant. Radiology. The pattern of a scaphoid fracture cannot be defined from a single radiological view. Neutral and ulnardeviated PA films do not show fracture lines well, because the axis of the scaphoid is flexed towards the beam and the tubercle overhangs the body. In these views, both surgical waist and dorsal sulcal fractures appear to be transverse and may be difficult to see except as a discontinuity in the radial cortex (Fig. 7). Fortunately, some views provide specific images for each type of fracture. To define all VOL. 80-B, NO. 2, MARCH 1998

222 J. P. COMPSON Fig. 6b Fig. 6a Proximal pole fracture: a) the anatomy; b) diagrams of an ulnar view to show the variation in size of the scapholunate joint (shaded) and the size of the fragment; and c) 45 oblique AP view showing the fracture line entering the radioscaphoid joint. Fig. 6c fractures, a series of scaphoid films should include 45 PA oblique (semipronated) and 45 AP oblique (semisupinated) views and one showing the long axis of the bone, such as an ulnar-deviated PA view with 20 angulation of the beam towards the elbow. Surgical waist fractures. A true lateral film and an ulnardeviated PA view with 20 angulation to the elbow lie closest to the plane of this fracture, but the lateral view (Fig. 4b) has variable value because of overlap by other carpal bones, although this projection is improved by taking it with the wrist extended. A transverse fracture line on an ulnar-deviated PA view with 20 angulation to the elbow is specific (Fig. 4c). A 45 PA (semipronated) oblique view sometimes gives good definition (Fig. 4d). Dorsal sulcal fractures. This pattern and any associated comminution are best shown on a 45 PA (semipronated) oblique view on which the fracture line runs from the dorsal apex of the ridge adjacent to the lunate, and passes obliquely (Fig. 5c). No other view shows this fracture line as clearly. Proximal pole fractures. These fractures may not be clearly visible unless they are displaced or have had time to develop sclerosis. The specific view is a 45 AP (semisupinated) oblique film (Fig. 6c), which will show the fracture crossing the radioscaphoid joint. A 45 PA (semipronated) oblique or a neutral or ulnar-deviated PA view shows the fracture well, only after often late displacement. The apparent size of the fragment and the angulation of the fracture on PA films depend on the relative size of the proximal pole, and also on the amount of anterior angulation of the scaphoid. Discussion Normally, two views of a bone are enough to define the plane of a fracture, but the scaphoid, with its significant natural twist at the waist and the overlap of other carpal bones, 23,24 needs at least four views for the full classification of the pattern of fracture. The recommended views are listed and discussed in Table I, with a neutral PA view included since it is often provided, although it does not help classification. It is difficult to classify scaphoid fractures from radiographs: careful study of dry cadaver bones or models is needed, and even then some method of three-dimensional reconstruction may be required in clinical practice. Other methods of imaging show displaced fractures and those which have collapsed after nonunion, 25,26 but these meth- THE JOURNAL OF BONE AND JOINT SURGERY

THE ANATOMY OF ACUTE SCAPHOID FRACTURES 223 Plain PA views of a) a waist fracture and b) a sulcal fracture to show the similarity on this projection. Fig. 7a Fig. 7b ods are not yet sensitive enough to define acute undisplaced fractures, while being of limited availability and too expensive for routine use. Classification based on plain radiographs has the problem that single views of different fracture patterns may look the same. An example is that a transverse fracture of the waist on Russe s classification 8,9 could actually be a type-2 or type-3 sulcal fracture or a fracture of the proximal pole with a large proximal fragment. The distinction requires a three-dimensional appreciation. The best single distinguishing feature between types of scaphoid fracture is the dorsal alignment of the fracture plane, and this cannot be seen or appreciated from a PA view. Some authors do not appear to understand fully the radiological changes which follow collapse of a scaphoid fracture: the same fracture has been classified differently before and after it has collapsed. 27 These changes are better understood when the fracture pattern is considered on a model or cadaver bone. Fractures of the distal pole or distal third reported in other classifications appear to be due to a similar radiological illusion. When a sulcal fracture collapses, flexion of the distal pole reduces the width of its image on a PA film and makes the fracture appear to be more distal. Excluding tubercle fractures, there were no true fractures of the distal pole or the distal third in the series. The two fracture lines which pass through the waist of the bone are transverse (surgical waist) or oblique (dorsal sulcal) to the long axis of the scaphoid. The sulcal fracture is probably much less stable than a surgical-waist fracture and is the only pattern which leads to the humpback scaphoid. Nakamura et al 26 describe volar and dorsal types of waist fracture, but do not relate these patterns to the osteology. If these two types, reported to have different patterns of collapse, correspond respectively to the waist and sulcal types described above, this would support the hypothesis that they have different natural histories. The comminution of scaphoid fractures has been noted, 28 but its extent is not well documented. In the 91 cases studied, undisplaced and displaced butterfly fragments were seen most commonly in dorsal sulcal fractures and to a less extent in those of the waist. In both types these were near the lateral apex of the ridge, on the palmar lateral side of the scaphoid. In general, butterfly fragments and comminution are most common on the compression side of a fracture, and this suggests that these types are not produced by hinging over the dorsal rim of the radius, which has been the usual explanation. 4,21,28-31 Such comminution in the area described would suggest compression rather than distraction. The comminuted area is also similar to the area of bone which is sometimes lost and needs replacement by an anterior wedge graft in some cases of longstanding nonunion. 16,25 Classifying scaphoid fractures in three dimensions may not appear to be easy. Any surgeon treating these difficult injuries, however, requires a thorough knowledge of the scaphoid, its anatomy, its anatomical variation, and its radiology. One of the challenges of orthopaedic surgery is the ability to think and operate in three dimensions. I would like to thank the Wishbone Fund of the British Orthopaedic Association, the Special Trustees of St Thomas Hospital for their financial support, and Professor F. W. Heatley for his encouragement and help in preparing this paper. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. References 1. Ripperger RR, Cooney WP, Linscheid RL. Distal pole scaphoid fractures. Orthop Trans 1980;4:18. 2. Prosser AJ, Brenkel IJ, Irvine GB. Articular fractures of the distal scaphoid. J Hand Surg [Br] 1988;13-B:87-91. 3. Mody BS, Bellliappa PP, Dias JJ, Barton NJ. Nonunion of fractures of the scaphoid tuberosity. J Bone Joint Surg [Br] 1993;75-B:423-5. 4. Mayfield JK, Gilula LA, Totty WG. Isolated carpal fractures. In: Gilula LA, ed. The traumatized hand and wrist. Philadelphia, etc: WB Saunders Company, 1992:252-3. 5. Compson JP, Waterman JK, Spencer JD. Dorsal avulsion fractures of the scaphoid: diagnostic implications and applied anatomy. J Hand Surg [Br] 1993;18:58-61. 6. Dickison JC, Shannon JG. Fractures of the carpal scaphoid in the Canadian army. Surg Gynaec Obstet 1944;79:225-39. 7. Stewart MJ. Fractures of the carpal navicular (scaphoid): a report of 436 cases. J Bone Joint Surg [Am] 1954;36-A:996-1006. 8. Böhler L, Trojan E, Jahna H. Behandlungsergebnisse von 734 frischen einfachen Brüchen des Kahnbeinkörpers der Hand. Reconstr Surg Traumatol 1954;II:86-111. 9. Russe O. Fracture of the carpal navicular: diagnosis, non-operative treatment, and operative treatment. J Bone Joint Surg [Am] 1960; 42-A:759-68. 10. Baumann JU, Campbell RD. Signficance of architectural types of fractures of the carpal scaphoid and relation to timing of treatment. J Trauma 1962;2:431-8. VOL. 80-B, NO. 2, MARCH 1998

224 J. P. COMPSON 11. Mazet R, Hohl M. Fractures of the carpal navicular: analysis of ninety-one cases and review of the literature. J Bone Joint Surg [Am] 1963;45-A:82-112. 12. Borgeskov S, Christiansen B, Kjaer A, Balslev I. Fractures of the carpal bones. Acta Orthop Scand 1966;37:276-87. 13. Eddeland A, Eiken O, Hellgren E, Ohlsson N-M. Fractures of the scaphoid. Scand J Plast Reconstr Surg 1975;9:234-9. 14. Leslie IJ, Dickson RA. The fractured carpal scaphoid: natural history and factors influencing outcome. J Bone Joint Surg [Br] 1981; 63-B:225-30. 15. Fisk GR. Injuries of the wrist. In: Wilson JN, ed. Watson-Jones Fractures and joint injuries. 6th ed. Edinburgh, etc: Churchill Livingstone, 1982;2:710-38. 16. Fernandez DL. A technique for anterior wedge-shaped grafts for scaphoid nonunions with carpal instability. J Hand Surg [Am] 1984; 9-A:733-7. 17. Dobyns JH, Linscheid RL. Fractures and dislocations of the wrist. In: Rockwood CA Jr, Green DP. Fractures in adults. Vol. 1. 2nd ed. Philadelphia, etc: JB Lippincott Company, 1984:411-509. 18. McLaughlin HL, Parkes JC. Fracture of the carpal navicular (scaphoid) bone: gradations in therapy based upon pathology. J Trauma 1969;9:311-9. 19. Herbert TJ, Fisher WE. Management of the fractured scaphoid using a new bone screw. J Bone Joint Surg [Br] 1984;66-B:114-23. 20. Cooney WP III, Dobyns JH, Linscheid RL. Nonunion of the scaphoid: analysis of the results from bone grafting. J Hand Surg [Am] 1980;5-A:343-54. 21. Weber ER. Biomechanical implications of scaphoid waist fractures. Clin Orthop 1980;149:83-9. 22. Szabo RM, Manske D. Displaced fractures of the scaphoid. Clin Orthop 1988;230:30-8. 23. Compson JP, Waterman JK, Heatley FW. The radiological anatomy of the scaphoid. Part 1: Osteology. J Hand Surg [Br] 1994;19: 183-7. 24. Compson JP, Waterman JK, Heatley FW. The radiological anatomy of the scaphoid. Part 2: Radiology. J Hand Surg [Br] 1997;22:8-15. 25. Belsole RJ, Hilbelink DR, Llewellyn JA, et al. Computed analyses of the pathomechanics of scaphoid waist nonunions. J Hand Surg [Am] 1991;16:899-906. 26. Nakamura R, Imaeda T, Horii E, Miura T, Hayakawa N. Analysis of scaphoid fracture displacement by three-dimensional computed tomography. J Hand Surg Am 1991;16:485-92. 27. Schernberg F, Elzein F, Gerard Y. Anatomo-radiological study of fractures of the carpal scaphoid bone: problems of abnormal callus. Rev Chir Orthop Reparatrice Appar Mot 1984;70Suppl2:55-63. 28. Saffar P. Carpal injuries: anatomy, radiology, current treatment. Paris: Springer-Verlag, 1990:91-112. 29. Weber ER, Chao EY. An experimental approach to the mechanism of scaphoid wrist fractures. J Hand Surg 1978;3:142-8. 30. Mayfield JK. Mechanism of carpal injuries. Clin Orthop 1980; 149:45-54. 31. Yanni D, Lieppins P, Laurence M. Fractures of the carpal scaphoid: a critical study of the standard splint. J Bone Joint Surg [Br] 1991; 73-B:600-2. 32. Bridgeman CF. Radiography of the carpal navicular bone. Med Rad Photo 1949;25:104-5. 33. Stecher WR. Roentgenography of the carpal navicular bone. Am J Roent 1937;37:704-5. 34. Ziter FM Jr. A modified view of the carpal navicular. Radiology 1973;108:706-7. THE JOURNAL OF BONE AND JOINT SURGERY