University of Groningen. Fracture of the distal radius Oskam, Jacob

Transcription

1 University of Groningen Fracture of the distal radius Oskam, Jacob IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record Publication date: 1999 Link to publication in University of Groningen/UMCG research database Citation for published version (APA): Oskam, J. (1999). Fracture of the distal radius: selected issues of epidemiology, classification and treatment Groningen: s.n. Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons). Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Downloaded from the University of Groningen/UMCG research database (Pure): For technical reasons the number of authors shown on this cover page is limited to 10 maximum. Download date:

2 FRACTURE OF THE DISTAL RADIUS SELECTED ISSUES OF EPIDEMIOLOGY, CLASSIFICATION, AND TREATMENT J. OSKAM

3 2

4 3 Lay out: Dr. W.F.M. Fritschy Druk: Print Partners Ipskamp BV, Enschede

5 4 RIJKSUNIVERSITEIT GRONINGEN FRACTURE OF THE DISTAL RADIUS SELECTED ISSUES OF EPIDEMIOLOGY, CLASSIFICATION AND TREATMENT PROEFSCHRIFT TER VERKRIJGING VAN HET DOCTORAAT IN DE MEDISCHE WETENSCHAPPEN AAN DE RIJKSUNIVERSITEIT GRONINGEN OP GEZAG VAN DE RECTOR MAGNIFICUS, DR. D.F.J. BOSSCHER, IN HET OPENBAAR TE VERDEDIGEN OP WOENSDAG 29 SEPTEMBER 1999 OM UUR DOOR JACOB OSKAM GEBOREN OP 18 DECEMBER 1963 TE ALKMAAR

6 5 PROMOTORES: PROF. DR. H.J. KLASEN PROF. DR. R. VAN SCHILFGAARDE REFERENT: DR. J. KINGMA ISBN

7 6

8 7 BEOORDELINGSCOMMISSIE: PROF. DR. H.J. TEN DUIS PROF. DR. P.A.M. VIERHOUT PROF. DR. R.P. ZWIERSTRA PARANIMFEN DRS. J.P. FRÖLKE IR. A. VAN DER WIEL

9 8 Acknowledgement The names of all the persons to whom I am very indebted are mentioned in this thesis. I gratefully thank all of them. Financial support was kindly received from: Mathys Medical Nederland BV Medi Nederland BV, Fabrikant verband - en therapeutische elastische kousen Orthomed BV Voor Didi, voor onze kinderen, en voor onze toekomst.

10 9 CONTENTS CHAPTER 1 Introduction CHAPTER 2 The Groningen Trauma Study...19 European Journal of Emergency Medicine 1994; 1: CHAPTER 3 Fracture of the Distal Forearm: epidemiological developments in the period Injury 1998; 29 : CHAPTER 4 Recognition of 10 distal radial fractures types by residents Submitted CHAPTER 5 The basic categories of the AO/ASIF s systems as a frame of reference for classifying distal radial fractures Submitted CHAPTER 6 Fractures of the distal radius and scaphoid Journal of Hand Surgery 1996; 21B: CHAPTER 7 Dorsal fracture-dislocation of the radiocarpal joint...65 Submitted CHAPTER 8 K-wire fixation for redislocated Colles fractures

11 10 Acta Orthopaedica Scandinavica 1997; 68: CHAPTER 9 Corrrective osteotomy for malunion of the distal radius.81 Archives of Orthopaedic and Trauma Surgery 1996; 115: CHAPTER 10 Ulnar shortening osteotomy after fracture of the distal radius.. 89 Archives of Orthopaedic and Trauma Surgery 1993; 112: CHAPTER 11 Summary and Conclusions CHAPTER 12 Samenvatting en Conclusies APPENDIX..113

12 11

13 12 CHAPTER 1 INTRODUCTION

14 13 HISTORICAL DEVELOPMENTS It was generally believed in the eighteenth century that a displacement of the wrist after injury was caused by a dislocation and not by a fracture. The notion that a distal radial fracture was a usual cause of wrist displacement gained popularity in the beginning of the nineteenth century. Autopsies or observations of patients with open fractures showed that a fracture of the distal radius was not rare at all. The famous French surgeon Dupuytren was most explicit in this regard (11). He wrote: Almost all authors who have written on dislocations of the wrist, have described as many as four kinds; and the only point in which these writers at all difer from each other is respecting the number... I have for a long time publicly thought that fractures of the carpal end of the radius are extremely common; that I had always found these supposed dislocations of the wrist turn out to be fractures; and that, in spite of all which has been said upon the subject, I have never met with or heard of, one single well authenticated and convincing case of the dislocation in question. Different types of distal radial fractures were distinguished by physical examination in the nineteenth century, by which in fact the first step in classification was made. Surgeons in several countries like Pouteau in France (21), Colles and Smith in Ireland (5,24), and Barton in the USA (1), described a specific type of distal radial fracture and their names are still connected to particular fractures. It seemed that there was not much controversy about the first choice of treatment and the ultimate functional outcome in those days. Abraham Colles reported that reduction was relatively easy, and good outcomes could usually be obtained after several weeks of immobilisation with specially designed lower arm splints (5). Dupuytren focussed the attention on the importance of early reduction, but also on technical problems to obtain adequate immobilisation (11): The surgeon should proceed to reduce these fractures (comminuted fractures of the distal radius, Oskam) as soon as possible, and that a misapprehension of the nature of the injury is always attended by impaired use and deformity of the fore-arm. If the fracture is not adequately immobilized it will dislocate again: If the tendency of the hand to move towards the radial side of the fore-arm be not counteracted, union will take place

15 14 whilst the bones are in this position; and deformity together with impaired power of rotating the fore-arm are the consequences. In some instances the displacement in question is so great, that this bone appears as if curved ; and many practitioners have been thus misled into the belief that there was dislocation of its carpal extremity. To prevent dislocation of the fracture fragments, Dupuytren developed a device in which the hand was kept in an abducted position: I have, by the above simple method succeeded to my entire satisfaction in curing the troublesome (unstable, Oskam) fractures, without any deformity or sacrifice of the rotatory motions of the fore-arm. After the introduction of radiography in 1895 it was soon appreciated that in the majority of wrist injuries the distal radius was fractured. It was also confirmed that malunion after healing of a distal radial fracture occurred more often than was usually thought. Due to radiology, surgeons could assess the quality of reduction, bone healing, and anatomical end results. It appeared that in a large number of patients a normal anatomical relationship was not obtained resulting in a malunion of the distal radius (12). A malunion of the distal radius was characterized by radial shortening, radial deviation, and dorsal angulation as Dupuytren had described earlier without the help of X-rays. Many surgeons thought that instability of the fracture was not the only cause of failure but that technical failure also played an important role (25). So, it was thought that malunion could be prevented in all instances by early adequate reduction and immobilisation. It was observed in the first decades of the twentieth century by several surgeons that the functional outcome of distal radial fractures was not favourable in patients with malunion. Disturbed anatomical relationships after fracture healing compromise wrist function and may cause pain, loss of wrist motion and diminished gripstrength of the hand. Particularly American surgeons became interested to treat anatomical deformations of the wrist, such as a prominent distal ulna due to radial shortening. This classical disorder was treated by resection of the distal ulnar head as was advocated by Darrach (9). Another technique to treat a prominent distal ulna was introduced by Milch. He developed ulnar shortening osteotomy as alternative for ulnar head resection (18). Apart from operations on the ulnar side of the wrist, operations for the malunited distal radius were also

16 15 promoted. For instance, Campbell developed radial corrective osteotomy to restore radiocarpal angulation and radial length (4). It became obvious that primary anatomical reduction could prevent late complications, and might prevent secondary reconstructive operations in many wrists. Nevertheless, most surgeons remained reluctant to perform primary operative fracture treatment. Primary surgery was only considered to be indicated in irreducible fractures, and in these cases open reduction was ususally followed by plaster immobilisation (25). Despite the fact that several progressive surgeons developed new operative techniques on the basis of sound clinical observations, these procedures were not generally accepted and, consequently, seldom mentioned in handbooks. Because these books usually reflected the conservative opinions of prestigious surgeons, all attention remained focussed mainly on closed reduction and plaster immobilisation (2,26). So, conservative treatment was undisputedly the treatment of first choice, irrespective of new insights on operative treatment for unstable fractures. PARADIGM SHIFTS IN TREATMENT OF DISTAL RADIAL FRACTURES Lambotte (16) and Matti (17) were two representatives of surgeons practising in the first decades of the twentieth century, who advocated open reduction and operative fixation if anatomical reduction could not be obtained by conservative methods. Their aggresive approach was based on the assumption that a strong relation exists between complete restoration of anatomy and wrist function. Lambotte and Matti challenged the leading opinions of conservative treatment protocols for wrist fractures. The interest in operative treatment remained modest despite their efforts. The quality of fixation material was rather poor in the years preceeding World War II, which at the time contributed to rather high rates of infection and redisplacement after operative treatment. This may well have contributed to the lack of succes of efforts to promote operative treatment of distal radial fractures.

17 16 It was confirmed by clinical studies published after World War II that the functional outcome after conservative treatment of a fracture of the distal radius was not always favourable. Gartland and Werley showed in 1951 that about 20% of patients complained of wrist pain and showed loss of wrist motion (13). As a consequence, attention was again focussed on the advantages of early restoration of anatomical relations to improve functional outcome. New clinical studies were undertaken in the nineteen-fifties to evaluate the results of osteosynthesis techniques in unstable fractures. Closed reduction and fixation with Kirschner wires became the first paradigm shift. It was the most popular method of operative treatment in the period (3,8,10,22). The approach of the distal radius was usually not open, because it was assumed that adequate reduction can be obtained by closed means. Open reduction was only indicated for irreducible fractures. Another paradigm shift in treatment occurred after 1970, largely due to the influences of the Arbeitsgemeinschaft für Osteosynthesefragen (AO) (8,19). The AO stressed that the basic principles of joint fracture treatment should be restoration of joint congruency, followed by rigid stabilisation of fractures to enable early practising and functional recovery. With respect to the wrist, restoration of anatomical relationships in the radiocarpal joint was considered to be a prerequisite for good functional recovery. The most appropriate procedure to reach this goal was open reduction and internal fixation (ORIF). To enable osteosynthesis of the distal radius with it s specific shape, a so called radius T- plate was designed to enable rigid fixation. Although the advantages of ORIF were theoretically promising it became gradually apparent that ORIF was not suitable for all fracture types (6,12,23). The best indications are presumably fracture-dislocations, like Barton s fracture, or unstable extra-articular volarly- or dorsally-displaced fractures (7,12). The principles of the AO and the assumed benefits of open reduction and internal fixation do not appear to be valuable for each distal radial fracture type. The most important disadvantages of ORIF are devitalization of fracture fragments and additional trauma to soft tissues which compromise fracture healing and may contribute to the occurence of infection. Most likely, closed fixation techniques are more

18 17 feasible to treat comminuted fractures, because the damage to soft tissues and to the vascularisation of bone fragments is less with a closed technique. Understandably, closed fixation techniques, like external fixation, became popular in the nineteen-eighties and constituted the third, and latest, paradigm shift (14). OUTLINE OF THE THESIS Quantitative epidemiological factors, like the incidence of wrist fractures or the incidence of particular fracture types, are likely to affect therapeutic decisions and surgical care. So, against the background of optimal clinical care, epidemiological information is an issue. Not only the incidence of distal radial fractures may be an interesting issue, but also aetiological or biological determinants of distal radial fractures are worth to be studied. For instance, the pattern of injury or the age of the patient are factors which eventually determine the fracture pattern in the distal radius. Little is known of epidemiological aspects of distal radial fractures. This lack of knowledge justifies further research because of the possible clinical consequences. Therefore, epidemiological issues are adressed in this thesis. There is a need for a sound classification system for distal radial fractures. Obviously, a correct diagnosis is required for choosing an appropriate treatment for a particular fracture type. Usually, a historical classification system with synonyms or eponyms is employed to reach this goal. Such historical systems do usually not contain well defined fracture items with the considerable risk that observer agreement is low. A more reliable classification system may be the AO classification system for distal radial fractures. It has been claimed by the AO foundation that this system contains all clinically meaningful fracture types and that it is treatment-based. Since it is important to choose the most appropriate treatment modality for a particular fracture, we have investigated the reliability of a historical classification system and the AO classification system. Conservative treatment can be performed in about 70 to 80 percent of distal radial fractures (7,12). If operative treatment is required, several technical approaches are available in most instances. There is no doubt that the adequate choice should

19 18 be made on the basis of sound evidence, preferably derived from prospective, randomized, controlled studies. In case of distal radial fractures, however, the number of such studies is regretfully low. As explanation may serve the observation that the number of different fracture types of the distal radius is more than thirty, in addition to the fact that each type can be treated with several operative techniques and that many of these fracture types have a very low incidence. Understandably, therefore, published observations as to success or failure of operative treatment of these fractures are mostly restricted to retrospective studies. Since we, too, have focussed on rare types of distal radial fractures, the studies presented in this thesis are retrospective by necessity. The first concern of this thesis regards the incidence and aetiological patterns of distal radial fractures in the Groningen population. No epidemiological studies of injury patterns have been published in The Netherlands. A unique achievement of the Department of Surgery of the University Hospital Groningen is that a trauma registration system with a database has been maintained since With that database, epidemiological aspects of a large group of trauma patients such as the incidence of wrist fractures in the Groningen area could be studied. Long term trends of incidence and injury patterns for all trauma patients treated at the University Hospital Groningen are presented in Chapter 2. More detailed incidence rates and injury patterns of patients with wrist fractures are discussed in Chapter 3. The second concern of this thesis is the reliability of classification of distal radial fractures. Many classification systems have been developed since 1950, but most systems are incomplete since they do not describe all possible types of distal radial fractures (12). A requisite for a proper classification system is that it covers the large number of fracture types which is estimated to be more than thirty, and that it is helpful in choosing an appropriate treatment. Therefore universal, treatment-based classification systems have been designed in the nineteeneighties. The universal system of the Mayo Clinics was introduced in the United States of America (7), while in Europe the universal system of the AO was propagated (20). The clinical value of these universal classification systems is not

20 19 known precisely. Recently, a validation study of the AO system showed that agreement between several observers is not very good (15). It appears that no reliable universal classification system is available to support clinical decision making. This could be the reason why many surgeons still use familiar, historical classification systems to deal with the complexity of distal radial fractures. We have confronted a group of surgical residents with a variety of X-rays with distal radial fractures, in order to decide upon the consistency of traditional classification systems. The results of the study are described in Chapter 4. Since the concept of a universal treatment-based classification system to support decision making is very attractive, we studied the applicability of the AO/ASIF classification system which is presently used by many in Europe. We determined observer agreement in a series of 124 distal radius fractures which was classified according to the AO/ASIF s system by two experienced observers. Chapter 5 contains the rates of agreement and a qualitative analysis to causes of disagreement. Finally, the third concern of this thesis is to assess the value of selected operative techniques which were performed to restore anatomic relationships in particular rare fracture types in order to investigate the postulated association between anatomy and function after healing of a distal radial fracture. In the ninetyseventies the concepts of the AO movement were also adopted by the trauma surgeons of the surgical department of the University Hospital Groningen, and as a result the treatment of distal radial fractures changed, too. Initially, most primary operations were performed in young patients with complex fracture types or high energetic wrist injuries. An example of such a complex wrist injury is a combination of fractures of the radius and scaphoid. The results of treatment of one of the largest published series is presented in Chapter 6. Another example of a complex, high-energetic wrist injury is dorsal fracture-dislocation of the radiocarpal joint. Although the most appropriate operation technique for this injury has yet not been esthablished, the best treatment is probably open reduction and internal fixation. Because the injury is rare, only small series or case reports have been reported so far. The experiences of the Groningen Department of Surgery

21 20 with a series of six patients, and the latest developments in the literature, are presented and discussed in Chapter 7. An issue in the treatment of distal radial fractures is how to deal with unstable fractures in mentally and physically healthy, elderly patients. The problem with this group of patients is that these patients commonly suffer from secondary osteoporosis. Osteoporosis causes technical difficulties during surgery because the mechanical properties of the distal radius do not allow firm fixation. Fractures in osteoporotic bones are frequently unstable after reduction and heal in malunion causing poor functional outcome. The classical approach in these patients is to start with conservative treatment. The X-rays usually show a satisfactory reduction of the fracture, but during follow up it appears that the fracture redislocates. Currently it is still not clear whether patients with redislocation may benefit from an operation in which wrist anatomy is restored. But if surgery is pursued several technical options may be used, among which closed reduction and Kirschner wire fixation is an attractive option because it is minimally invasive. Our experiences with closed reduction and Kirschner wire fixation after redislocation are presented in Chapter 8. If a fracture has healed in malunion the anatomic relationship of the radius and ulna can be restored secundarily by a corrective osteotomy. Corrective osteotomies of the wrist have been performed in the Groningen surgical department since nineteen-eighty. Most of these reconstructions took place in young patients because of poor wrist function after conservative treatment. The preferred types of reconstructions were ulnar shortening osteotomy and radial corrective osteotomy. A surgical audit to evaluate the clinical outcome and suggestions to improve surgical techniques are described in Chapters 9 and 10. REFERENCES 1. Barton JR. Views and treatment of an important injury of the wrist. The Medical Examiner 1838; 1: Boehler L. Die Techniek der Knochenbruchbehandlung, Verlag Wilhelm Maudrich, Wien 1932, pp

22 21 3. Boehler L. Die techniek der Knochenbruchbehandlung, Erganzungsband, Verlag Wilhelm Maudrich, Wien 1963, pp Campbell WC. Malunited Colles fractures. JAMA 1937; 109: Colles A. On the fracture of the carpal extremity of the radius. Edinburgh Med J 1814; 10: Cooney W P, Linscheid R L, Dobyns J H. External pin fixation for unstable Colles fracture. J Bone Joint Surg 1979; 61A: Cooney WP and Saffar P. Fractures of the Distal Radius. Martin Dunitz, London Danis R. Theorie et pratique de l osteosynthese. Masson et Cie, Paris Darrach W. Partial excision of lower shaft Ulna for deformity following Colles fracture. Ann Surg 1913; 57: DePalma AF. The management of fractures and dislocations, WB Saunders Comp Philadelphia 1959, pp Dupuytren G. On the injuries and diseases of bones, selections from the collected edition of the clinical lectures. Translation F LeGros Clark, Sydenham Society, London Fernandez DL and Jupiter JB, eds. Fractures of the distal radius, Springer-Verlag New York Gartland J J, Werley C W. Evaluation of healed Colles fractures. J Bone Joint Surg 1951; 33A: Haas J L, Caffiniere de la J Y. Fixation of distal radial fractures: Intramedullary pinning versus external fixation. In Fractures of the distal radius (Eds Saffar P, Cooney WP). Martin Dunitz London 1995; 27: Kreder HJ, Hanel DP, McKee M, Jupiter J, McGillivery G, Swiontkowski MF. Consistency of AO fracture classification for the distal radius. J Bone Joint Surg 1996; 78B: Lambotte A. L intervention operatoire dans les fractures, Lamertin, Bruxelles Matti H. Die Knochenbruche und ihre Behandlung, Springer Verlag, Berlin 1931, pp Milch H. Cuff resection of the ulna for malunited Colles fracture. J Bone Joint Surg 1941; 23A: Mueller ME, Allgoewer M, Willenegger H. Manual der Osteosynthese AO-technik, Springer Verlag, Berlin Mueller ME. Two pamphlets of the Comprehensive Classification of Fractures, edition AO/ASIF Documentation Center Davos, Switzerland Pouteau C. Oeuvres posthumes. PhD Pierres, Paris 1784, tome second. 22. Rayhack J M. The history and evolution of percutaneous pinning of displaced distal radius fractures. Orthop Clin North Am 1993; 24: Riis J, Fruensgaard S. Treatment of unstable Colles fracture by external fixation. J Hand Surg 1989; 14B:

23 Smith RW. A treatise on fractures in the vicinity of joints and on certain forms of accidental and congenital dislocations. Hodges and Smith, Dublin Speed K. Fractures and Dislocations, Lea & Febiger, Philadelphia 1928, pp Watson-Jones R. Fractures and other bone and joint injuries, E&S Livingstone, Edinburgh 1940, pp

24 23 CHAPTER 2 THE GRONINGEN TRAUMA STUDY: INJURY PATTERNS IN A DUTCH TRAUMA CENTRE J Oskam, J Kingma, H J Klasen Department of Surgery, University Hospital Groningen Groningen, The Netherlands European Journal of Emergency Medicine, 1994; 1:

25 24 A trauma registry is an essential part of trauma care, and can be employed to evaluate injury characteristics (5,9). Registration systems give access to information on past injury patterns, and may provide directions for public health and trauma care management (25). Previous studies of large trauma populations mainly involved analyses of trauma mortality, because death is often best documented (1,5). In contrast, only one large Northern American study has also addressed characteristics of non-fatal injury in a large group of trauma victims (10). To our knowledge, injury patterns describing the entire spectrum of trauma care in a European trauma centre have not yet been reported. At the University Hospital of Groningen, a computerized trauma registry has been established since 1970, in which all primary visits irrespective of severity grade have been recorded. Thus, long-term information on trends of a large group of trauma victims (n=245,251) is currently available. This retrospective study was undertaken to analyse specific characteristics of injury patterns in our hospital. We will assess how this information was used to support trauma care management. Furthermore, the usefulness and future implications of trauma registry in the Netherlands will be discussed. MATERIAL AND METHODS The University Hospital of Groningen is a 1050-bed centre and is situated in the north of The Netherlands. The A&E department is freely accessible and maintains a 24-hour service. Since 1970, all trauma visits have been recorded on a standardized chart. Each case record comprises patient identification, external cause of injury, co-morbidity, trauma diagnoses, therapeutic procedures, complications, length of hospital stay and other treatment characteristics. All patient maps were reviewed and completed by staff traumatologists. The input of the data was performed by trained personnel. Diagnoses of the injuries and the external cause were recorded according to ICD- 8 untill 1981, and according to ICD-9-CM from In order to analyse injury causes for the 24-year period, all ICD-9 E-codes were converted into ICD-8

26 25 categories (6). The aetiologic category Fall was defined as an unintended fall not restricted to a certain height, and not related to work. Traffic was defined as accidents that happened in an area reserved for traffic. The victims are either drivers or passengers of cars, motorcycles, mopeds and bicycles and pedestrians. Violence was defined as incidents purposely inflicted, irrespective of the use of arms. In this category, however, a separate group of firearm injuries was selected. The group of Sports and unspecified injuries consisted of accidents which took place while a person was at home or engaged in some kind of sports activity. Information on age, sex, trauma visit, hospital admission and hospital mortality were collected from the database. Mortality was defined as death occurring during either hospital admission or hospital stay. To describe the influence of age on injury cause, a subdivision into seven age groups of 10 years each, and a group of 70 years and older was made. To analyse sex differences, the Male to Female ratio (M/F) was used. The case fatality ratio (CFR) per cause category was employed to describe the distribution of fatal and non-fatal injury. So, the relative contribution to mortality of each injury cause could be assessed. Cause category % ICD 8 E-code Fall 27 E887 Sports/unspecified 26 E929 Traffic 19 E Cutting/piercing instr. 10 E920 Violence 4 E Late effect 4 E946 Falling object 3 E926 Machinery 3 E928 Animal 1 E906 Rest (drowning, etc.) 3

27 26 Table 1. Distribution of the major causes of trauma visit (n=245,251) in the period RESULTS Injury causes and trends During the 24-year period, 245,251 primary trauma visits were registered. The mean annual incidence rate was 10,219 patients. In total, 29,430 hospital admissions (12%) occurred with a mean annual rate of 1,226. Table 1 shows that fall was the main injury cause, while almost a quarter of all patients suffered from injuries that resulted from some kind of leisure activity (26%). Remarkably, causes related to occupation, like machinery (3%) or falling objects (3%), were observed in a minority of cases. Injuries due to violence concerned only 4% of the primary visits, thus, not being a major contributor to trauma morbidity in Groningen. The incidence trends of the principal trauma causes are shown in Figure 1. The total incidence of trauma visits remained approximately 10,000 victims per year. However, a clear alteration in the distribution per cause category can be found. Although, both traffic and accidental fall remained two major trauma causes, a decrease in incidence can be observed. During the past 10 years, sharp objects and violence were causes of increasing importance. Trends in traffic accidents A traffic accident was the cause of injury in 19% of trauma patients. Figure 2 shows a comparison of the incidence of hospitalized patients treated at the University Hospital Groningen and all Dutch hospitals (Central Bureau of Statistics). It appears, that the local traffic injury pattern resembles the national pattern, while the overall incidence is currently decreasing. In Figure 3, the incidence rates for specific types of traffic accidents are shown. It can be seen that the incidence of injuries due to car accidents has decreased, while the incidence of injuries due to motor cycle accidents remained stable. The incidence of pedestrian injury decreased, but the sharp increase of bicycle injuries upto 939 patients is remarkable. Bicycle and pedestrian injuries comprised 66% of all traffic

28 27 casualties, while 52% of traffic mortality occurred in these traffic participants. Altogether, pedestrians and bicyclists constituted a major proportion of trauma patients, with children and elderly being most at risk

29 28 number year Accidental fall Traffic Cutting/ piercing instr. Violence Figure 1. Annual incidence rates of trauma visits to the University Hospital Groningen in the period number year Netherlands UHG Figure 2. Incidence of hospitalized trauma patients treated at the University Hospital Groningen (UHG) and all hospitals in the Netherlands. number year Car Motorcycle Bicycle Pedestrian Figure 3. Annual incidence rates of trauma visits for several subgroups of traffic accidents

30 29 Demographic characteristics An analysis of sex and age patterns revealed that the mean Male to Female (M/F) ratio was 1.8 (157,245 men, and 88,006 women). The highest incidence rate per age and sex group could be attributed to men of years. This particular pattern could also be observed in the cause categories traffic and violence. However, in accidental fall the highest incidence was found in both children (0-9 years) and women above 70 years. It also appeared that the local pattern did not differ from the national age and sex characteristics. The general demographic picture reflects a pattern in which young men in particular are susceptible to traumatic injury. Cause Number % CFR Traffic Fall Sports/ unspecified Violence Falling objects Machinery Rest Total Table 2. Survey of the principal causes of trauma mortality (n=1138) during the period at the University Hospital Groningen (Note: The Case Fatality Ratio (CFR) is the quotient of the number of deceased victims and all treated victims per injury cause) Trauma mortality In total, 1138 patients died soon after or during hospital admission, resulting in a mean of 48 deaths per year. The mortality rate of trauma patients was 0.5%, while the mortality rate of hospitalized patients was 4%. Not surprisingly, mortality was highest in the age group above 70 years, and was caused by both accidental fall and traffic accidents.

31 30 It can be seen in Table 2 that traffic contributed most to mortality. Violence was not a major cause of death, because homicide comprised only 3% of mortality. Remarkably, only 165 firearm injuries were registered during the 24-year period of observation. Traffic accidents caused the relatively most fatal injuries, while violence was the second most lethal injury cause. The mean CFR of all motor vehicle injuries was 40, which means that these injuries were the most dangerous. DISCUSSION The data of this study were retrieved from the Groningen Trauma Registry, a system which was computerized from the onset in 1970, with the WHO classification of diseases (ICD-CM) as a central tenet. Because we record all trauma visits, we have the opportunity of assessing the entire workload of a large trauma centre. The Groningen Trauma Registry has therefore been used as an information supply for the development of trauma care policy (8). However, since in present study the data of a single hospital have been used, it could be argued that selection might have biased the injury patterns. Therefore, the question arises whether single hospital data are appropriate for the purpose of trauma care management. A strong argument against the issue of selection is the observation of a striking resemblance between Groningen and Dutch injury trends. Furthermore, the local and national demographic patterns also appeared to be similar. Apart from the above mentioned resemblance of local and national trauma care statistics, the contemporary organization of the Dutch trauma care system might also explain why selection is probably of less importance. Usually, most trauma victims in The Netherlands are transported to the nearest hospital in the area where the accident took place. This means that the Dutch trauma care system is not regionalized and that almost all patients with thoracic, abdominal and extremity injuries are being treated in local hospitals (7,8). Only patients with severe head and spine injuries, and critically ill patients with secondary organ failure following severe trauma are reffered to the Groningen trauma clinic. Since, in our hospital the yearly incidence

32 31 is 10,000 victims on average, and referred patients are only a small fraction of the total trauma incidence, the injury patterns presently under study will probably not be biased. Thus, it seems that the present hospital population might be a reasonable sample of the Dutch trauma care system. From the above point of view, we think employing information of a hospital-based system for a broader perspective than just clinic management is justified. It has been discussed previously, that injury dynamics of a distinct population are determined by both time-related developments and geographic influences (26). Currently, the role of a difference in geographic area can be well demonstrated. For example, the Groningen inhospital mortality rate of 4% is much lower than the reported rate of 9% in the Major Trauma Outcome Study (MTOS) (5). The MTOS is a well known multicentre study, describing several aspects of trauma care in the USA. The variation in mortality rate undoubtedly reflects a difference in trauma severity. The observed contrast might be better understood by what Eastman called inclusive and exclusive trauma care (9). Most trauma centers of the US trauma care system provide exclusive care, which is mainly directed at critically injured patients. In Groningen, however, trauma care comprises all types and severities of injury, and this can be considered inclusive trauma care. The fact that 88% of patients were treated in an outpatient setting, and 99,5% suffered from non-fatal injury, underlines the inclusive character of trauma care in Groningen. Apart from differences, several similarities between the Groningen trauma population and other trauma populations can be discerned. For example, the demographic pattern is characterized by what can be called the young male peak. This peak is nearly universal (2,10,11,12), and is thought to result from the particularly high risk behavior of young men. The young male peak is present in all cause categories except in accidental fall, in which boys (0-9 years) and women older than 70 years are the major at-risk groups. These findings are in agreement with studies from the USA and Scandinavia (3,20,22). Furthermore, most deaths occurred in the elderly and resulted from traffic injuries and falls, a pattern which has also been observed in other Western societies (21,23). Traffic injuries caused most trauma deaths, in particular in young men between and years, a pattern which has been described elsewhere (4,5,12). The present case fatality

33 32 ratio of motor vehicle injury was 40 and resembles closely the CFR of 41 in the USA (24). In summary, not withstanding the differences, a striking similarity with Northern American, Scandinavian, and British trauma populations appears to be present. The strength of trauma registry systems lies in the ability to obtain information for the management and organization of trauma care (9). The costs of registration, however, are only justified if they are incorporated in a trauma system which is not only directed to the clinical level, but is also aimed at preventative measures (25). Currently, the Groningen Trauma Registry is not just being used for hospital management, but also to support preventative programmes. A good example of such a programme is a local project to prevent assaultive injury. In several papers, we described the role of alcohol consumption (15,16), the effects of increased waepon use (17,18), and a model to study the streams of violence victims (19). As a result, a program to prevent violence injuries was developed in association with police, county, and public health officers. The outcome of the program will be evaluated, also with the help of the trauma registry. Another application of the trauma registry is a national study of bicycle injuries. This study has been initiated by the Dutch Association of Safe Traffic (Veilig Verkeer Nederland) on the basis of our observation of an increase in trauma visits of bicyclists and pedestrians. This rather surprising information could be retrieved because the registry is well equiped to monitor long-term trends. Because the national and Groningen incidence patterns of traffic injury seem to be related closely, it was suggested that this trend could also exist in other parts of The Netherlands. As a result, a national project to study bicycle injuries was started in order to obtain more insights into the aetiology and prevention. Investigations of the underlying mechanisms of occupational-related accidents are scarce. Since, approximately 35-40% of incidence comprises minor trauma due to leisure-related or occupational accidents, we initiated new research in the field of non-fatal injury. An example is our recent observation that industrial hand injuries occurred mainly to men older than 50 years (13,14). It was discussed that factors like visual acuity might play a crucial role and, thus, this issue is now under study. Although, complete solutions are not to be expected, it must be realized that the role

34 33 of trauma registry in this kind of research is mainly description of injury patterns, and monitoring of the effects following introduction of preventative measures. In line with Trunkey's arguments (25), future trauma research must be directed to the evaluation of non-fatal injuries, in order to become more efficient in trauma care. Trauma registries must be used as tools to reach this goal. The present study shows that hospital registries may supply the data for analysis of non-fatal injury. Because the Groningen Trauma Registry records inclusive trauma care since 1970, changing trauma trends across the whole spectrum of injury severity could be assessed. The information thus derived served as a starting point for community trauma care programmes. Therefore, we conclude that trauma registration is an important instrument to monitor specific characteristics of patients, as well as causes of accidents, providing a tool for trauma management and public decision making. REFERENCES 1. Burns CM. The 1990 Fraser Gurd lecture: A Canadian trauma registry system nine years experience. J Trauma 1991; 31: Carlsson GS, Svardsudd K, Carlsson S, Tibblin G. A study of injuries during life in three male populations. J Trauma 1986; 26: Centers for Disease Control. Childhood injuries in the United States. Am J Dis Child 1990; 144: Cesare J, Morgan AS, Felice PR, Edge V. Characteristics of blunt and personal violent injuries. J Trauma 1990; 30: Champion HR, Copes WS, Sacco WJ, Lawnick MM, Keast SL, Bain LW, Flanagan ME, Frey CF. The Major Trauma Outcome Study: Establishing national norms for trauma care. J Trauma 1990; 30: Commission on Professional and Hospital Activities. The International Classification of Diseases. Nineth Revision-Clinical Modification. Ann Arbor,MI: WHO Draaisma JMTh, De Haan AFJ, Goris RJA Preventable trauma deaths in the Netherlands - A prospective multicenter study. J Trauma 1989; 29: Dutch College for Hospital Supplies. Report Workgroup on Trauma Care. Utrecht, The Netherlands: Ministry of Health Eastman AB. Blood in our streets; the status and evolution of trauma care systems. Arch Surg

35 ; 127: Fife D, Barancik JI, Chatterjee BF Northeastern Ohio Trauma Study: II. Injury rates by age, sex, and cause. Am J Public Health 1984; 74: Honkanen R, Koivumaa H, Smith G. Males as a high risk group for trauma: The Finnish experience. J Trauma 1990; 30: Guirguis EM, Hong C, Liu D, Watters JM, Baillie F, McIntyre RW. Trauma outcome analysis of two Canadian Centres using the TRISS method. J Trauma 1990; 30: Jonge de JJ, Kingma J, Lei van der B, Klasen HJ. Phalangeal fractures of the hand; an analysis of gender and age-related incidence and aetiology. J Hand Surg 1994; 19B: Jonge de JJ, Kingma J, Lei van der B, Klasen HJ. Fractures of the metacarpals. A retrospective analysis of incidence and aetiology and a review of the English-language literature. Injury 1994; 25: Kingma J, Klasen HJ. Alcohol consumption in victims of traffic accidents; car and motorcycle drivers in the period Tijdschrift Alcohol en Drugs 1994; 19: Kingma J, Oskam J, Klaver A, Klasen HJ. Alcohol consumption in victims of assault: A trend study of the period Tijdschrift Alcohol en Drugs 1992; 18: Kingma J, Oskam J, Eijken T, Klasen HJ. Use of weapons is increasing in Groningen. Tijdschrift Samenleving Criminaliteitspreventie 1994; 8: Oskam J, Kingma J, Klasen HJ. The use of objects and waepons in violental injuries. Tijdschrift Criminologie 1994; 36: Oskam J, Kingma J, Klasen HJ. The treatment of victims of violence in a hospital. Justitiele Verkenningen 1994; 20: Peclet MH, Newman KD, Eichelberger MR, Gotschall S, Guzzetta PC, Anderson KD, Garcia VF, Randolph JG, Bowman LM. Patterns of injury in children. J Ped Surg 1990; 25: Ryynanen OP, Kivela SL, Honkanen R, Laippala P, Soini P. Incidence of falling injuries leading to medical treatment in the elderly. Public Health 1991; 105: Sjogren H, Bjornstig U. Unintentional injuries among elderly people: Incidence, causes, severity, and costs. Accid Anal Prev 1978; 21: Smith DP, Enderson BL, Maull KI. Trauma in the elderly: Determinants of outcome. South Med J 1990; 83: Soderstrom CA, Birschbach JM, Dischinger PC. Injured drivers and alcohol use: Culpability, convictions and pre- and post-crash driving history. J Trauma 1990; 30: Trunkey DD. Future shock. Arch Surg 1992; 127: Waller JA. Methodologic issues in hospital-based injury research.j Trauma 1988;28:

36 35

37 36

38 37 CHAPTER 3 FRACTURE OF THE DISTAL FOREARM : EPIDEMIOLOGICAL DEVELOPMENTS IN THE PERIOD J Oskam, J Kingma, H J Klasen. Department of Surgery, University Hospital Groningen Groningen, The Netherlands Injury 1998; 29:

39 38 In spite of the fact that fractures of the distal forearm are often diagnosed, epidemiological studies to distal forearm fractures are scarce. A search in Medline over the period revealed only two epidemiological studies that dealt exclusively with distal forearm fractures. In a study from Oslo in Norway, age and gender specific incidences of patients older than 20 years were described (1). Another study from Malmo in Sweden reported on the incidence across the whole lifespan (2). In this study, a comparison was performed with historical data from the period and in the same hospital. It was concluded, that the incidence of distal forearm fractures was increased significantly, and that people older than 60 years were the highest risk group. Furthermore, the authors extrapolated the results to the future, and they forecasted that the incidence of distal forearm fracture would increase further. However, this extrapolation has been based on the assumption that the observed trend would continue to increase in the nineteen-eighties. MATERIAL AND METHODS The purpose of the present study was to investigate the longterm incidence rate in the period , enabling us to evaluate the forecasted trend. Another concern was to analyze aetiologic factors and the clinical workload of distal forearm fractures in our department. Since 1970, a trauma registry has been established at our institute and information on a large group of patients is nowadays available (3). The records of all patients with a fracture of the distal forearm treated from 1 January 1971 through 31 December 1995 were reviewed. All patients were seen primarily at the Emergency and Accident unit of our hospital, which is a 1,056-bed centre with the largest trauma centre in the northern region of the Netherlands. The unit holds a 24-hour service and is the only institute in the area which is accesible for all kinds of trauma. The proportion of trauma admissions to our hospital in the catchment area is 94%. The admission data were obtained from the patient s history, clinical and radiological assessment at the time of first consultation. All trauma visits have

40 39 been recorded on a standardized chart by the attending physician, and were checked by trauma-surgeons. Fracture of the distal forearm was defined according to the N-code of the International Classification of Diseases (N 813.0). The Supplementary Classification of External Causes of Injury (E-code) was used to study the aetiology of the injury (3). The category Traffic was defined as an accident in which at least one vehicle is involved and that happened in an public area (E ). The vicitims are either drivers or passengers of cars, motorcycles, mopeds, bicycles, or pedestrians. The group of Sports & leisure consisted of accidents which took place while a person was engaged in some sport or leisure activity. Information on age, sex, injury cause, and hospital admission was collected from the trauma data base. To describe the distribution of age, a subdivision into eight age groups of ten years each, and a group of 80 years and older was made. The population was defined as the total of inhabitants living in the catchment area of the hospital during the period of observation. The populational data were delivered by the Central Bureau of Statistics (CBS) of the Netherlands. The annual incidence of distal forearm fractures per 10,000 inhabitants was employed as a measure of disease frequency rate. The nominator data were the number of patients, while the denominator was the number of inhabitants in the catchment area (incidence = number patients / population x 10,000). Apart from the longterm incidence, the age- and sex-specific incidence, and the incidence for several injury causes were computed. The absolute number of hospital admission was used to investigate the workload in our hospital. RESULTS Incidence rates It appeared that 8,567 distal forearm fractures out of a total of 256,431 trauma visits (3%) were recorded during the period of observation. The mean incidence rate across the whole population increased in the period , while a gradual decrease can be observed in the period (see Table 1). Turning to riskgroups, the highest age-specific incidence rate ocurred in the group older

41 40 than 79 years throughout the whole period of observation. The age group of 0 to 9 years was the second greatest risk group at the end of 1995, due to an increasing incidence rate. In contrast, a decrease was observed in the age groups of 50 to 79 years. While, the incidence rate remained rather stable in the age groups of 20 to 49 years. In respect to gender characteristics, the mean male to female ratio across the whole lifespan showed a slight predominance for women (M/F 1:1.4). Under the age of 40 years almost no sex predominance was present. Whereas, above 50 years a clear turning point in male/female ratio can be noticed, and more women than men were treated (see Table 1). Age group (years) Period (per 10,000) Male/female ratio : : : : : : : : 5! : 4 mean incidence per 10, : 1.4 population * 845, , , , ,197 Table 1. Incidence rate per 10,000 inhabitants of distal forearm fractures (n= 8567) per age group in the period 1971 to Generally, a gradual decrease in incidence rate can be observed. If a subdivision of incidence is made in a low (10-29 per 10,000), an intermediate (30-59 per 10,000), and a high risk group ( per 10,000), the age groups of 0-9, years, and above 70 years are most at risk. Legend:* total population in adherence area in absolute number.

42 41 Aetiology Accidental fall was the major injury cause across the whole lifespan ( see Table 2). Despite the fact that the incidence rate of accidental fall decreased constantly from 32 in 1971 to 22 per 10,000 in Furthermore, it appeared that the distribution of the three major causes remained stable in 25 years. Another frequently observed injury cause was sport & leisure, especially in the age groups of 0 to 9, and 10 to 19 years. However, the longterm trends of both sport & leisure, and traffic were stable with 7 per 10,000. Age group Accidental fall Sport & Leisure Traffic Other Total ! Table 2. Distribution of the main causes of distal forearm fractures per age group in incidence rate per 10,000 inhabitants. The mean over the whole period of 25 years is displayed. In-patient treatment In-patient treatment was performed in 886 out of 8,567 patients (10%). The pattern of injury cause in case of inpatient treatment was respectively accidental fall in 50%, traffic in 29%, and sports & leisure in 15% of cases. It is shown in Table 3, that per 5-year period the number of hospital admission changed with both increases and decreases. However, it seems clearly that on the long term the

43 42 proportion of in-patient treatment increased almost two-fold from 6 percent in 1971 to 14 percent in In respect to age, the greatest rise can be seen in the age group of 0 to 9 years. However, a marked increase also ocurred in the age groups of 0 to 49 years. Although, in the age group of years an increase can be observed, it appears that the longterm increase of hospital admission can largely be attributed to patients younger than 50 years. Age group (years) Period Total inpatient treatment fractures > Total inpatient Total fractures Table 3. An oversight of the clinical workload in the period across the whole lifespan is shown. Instead of the incidence, the absolute numbers of patients are displayed. Both the total of all treated patients and the total of only inpatient treatment are shown in the last two vertical columns. The trends of inpatient treartment per age group are shown, horizontally.

44 43 DISCUSSION This study shows, that the incidence rate of distal forearm fracture increased in the period 1971 to 1980, but a clear decrease occurred from 1981 to Interestingly, the increase in the nineteen-seventies is similar to the incidence trends that were reported by Scandinavian authors (1, 2). Bengner et al., compared the incidence rates of the period with those of 1981and It was predicted that an ongoing increase would occur in the years thereafter. In contrast, we presently found a decrease in incidence rate. Apparently, an extrapolation of observations based on historical data is not always warranted, and a forecasted trend may not appear, actually. It seems, that trends might better be monitored longitudinally to study epidemiological developments. Turning to risk groups, it appeared that patients older than 79 years remained the highest risk group during 25 years. It has been reported elsewhere that people older than 60 years were the main risk group (2). So, a shift from people towards the oldest patients in the lifespan might have taken place. The clinical context of this change may be noteworthy, because biological factors like osteoporosis and comorbidity may interfere with fracture healing, and with functional outcome in older patients. Furthermore, it seemed that the number of distal forearm fractures in children has increased enormously. An increase in fractures in children may pose the clinician more often to specific problems like growth disturbances in the wrist. Paradoxically, while the incidence rates decreased the rate of hospital admission increased two fold since This rise can not be explained by alterations in the extent of the population in the area of adherence, because the population grew with only 0.5 % in 25 years. Therefore, we think that policy changes in the nineteen-eighties most likely caused the observed change. Firstly, since 1980 reduction of displaced fractures in children was no longer performed under local but under general anaesthesia causing an increase in registration of inpatient treatment. Secondly, the observed rise in patients of 30 through 50 years may

45 44 have caused an increase in clinical treatment. Because these patients are more often operated on. During the eighties new insights gained popularity in the field of wrist surgery, and the aim was to strive for optimal anatomical results. As a consequence, more indications for primary operative treatment were also employed in our department (4, 5). REFERENCES 1. Falch JA. Epidemiology of fracture of the distal forearm in Oslo, Norway. Acta Orthop Scand 1983; 54: Bengner U, Johnell O. Increasing incidence of forearm fractures. Acta Othop Scand 1985; 56: Oskam J, Kingma J, Klasen HJ. The Groningen Trauma Study. Eur J Emerg Med 1994; 1: Oskam J, Bongers K, Karthaus AJM, Frima AJ, Klasen HJ. Corrective osteotomy for malunion of the distal radius. Arch Orthop Trauma Surg 1996; 115:

46 45

47 46 CHAPTER 4 RECOGNITION OF 10 DISTAL RADIAL FRACTURE TYPES BY RESIDENTS. J Oskam, J Kingma, A J M Karthaus, H J Klasen. Departments of Surgery, University Hospital Groningen and Deventer Hospital the Netherlands

48 47 Submitted Patients with a distal radial fracture are commonly treated by physicians with little experience. Since, many specific types of distal radial fractures have been described, and different therapeutic regimens can be choosen, there is a need for a classification method which is easy to handle. A search in Medline showed that at least thirteen classification systems have been reported since It seems, that no classification system has been proven to be superior, and a generally accepted frame of reference for inexperienced physicians to classify distal radial fractures is still lacking. Although, universal classification systems has been introduced, it appears not to be uncommon that fractures are usually provided with labels referring to either the first author describing the particular fracture type (synonyms), or referring to an injury mechanism (eponyms) (2). Given the fact that a universal classification system is often not being used, a question to be answered is how the base-line of the clinician s recognition of distal radial fractures can be established. Knowledge about the performance of recognizing the different fracture types may reflect the actual classification ability, and may serve as a starting point to develope useful programs to teach the essence of distal radial fractures (7). In addition, a strategy may be designed how modern, treatment-based classification systems might be introduced to clinicians (5). The purpose of the present study is to investigate the verbal and visual recognition of 10 commonly cited distal radial fractures (1,10). MATERIAL AND METHODS Participants were 30 surgical residents from five teaching hospitals, who had on the average 2 years clinical experience (range 1-4 years). The residents participated in a test in which they were asked to assess a series of 10 different distal radial fracture types. The series of fracture types was developed with an increasing level of complexity. Several simple fractures and some specific intraarticular fractures, fracture-dislocation, and combination fractures were included. The series of 10 fracture types consisted of respectively: Colles, Smith s, distal

49 48 forearm, a combination of radius & scaphoid, radial styloid process, dorsal Barton s, volar Barton s, pilon, chauffeur s, and lunate load fracture (1, 4, 8, 10). The test consisted of two parts; a verbal and a visual part with 10 items each. Each verbal item consisted of a description of a distal radial fracture. At least two relevant distinctive features of the specific fracture type were used for the description of each verbal item (see Appendix). In each verbal item the subject was asked whether he recognized the particular fracture in the description. The 10 questions were printed on one sheat and each verbal statement was followed by a question about whether the clinician recognized the particular fracture in a yes or no format. The 10 corresponding visual items contained an X-ray (AP and lateral projection) of each fracture type. Only those X- rays were included in which there was complete agreement between the authors about the type of fracture and the clearness of the X-ray. Each fracture type on an X-ray corresponded with the concomitant verbal item. The administration of the tasks was in a random order. The subjects were asked to write down the name (diagnosis) of the fracture type on a sheat. In case they didn t know an exact label of the fracture type, they were asked to write down the relevant distinctive features of the particular fracture. Before test administration, two introductionary examples were given for each kind of test to accustom the subjects to the questions that were posed. The verbal test was administered at first, followed by the visual test. To avoid bias due to repeated measurement the verbal and visual recognition items were administered in a random counterbalanced order for each test (9). The administration of the 10 verbal and visual items together took about 45 minutes. For each verbal item affirmative responses ( yes ) were scored as 1, and no responses were scored as 0. For each visual item, a correct respons (score 1) was defined as either a correct diagnostic label, or a description in which at least two distinct features from the corresponding fracture description were correctly used. A zero score was given in all other instances. The maximal total score per participant ranged from 0 to 10 for both the verbal and visual counterparts. The percentage of positive ( yes ) answered verbal questions, and correctly diagnosed X-rays was computed per fracture type for the whole group of subjects.

50 49 By definition, the criterion of a score more than 80 percent correct responses per item for both verbal and visual recognition was choosen as a standard of adequately diagnosing the particular fracture (9). The binomial (Z) test was used to test the percentage of correct responses against the criterion of 80 percent for each fracture type. The Spearman rank correlation coefficient (Rs) was employed to show the extent of agreement between the rank order of scores in the verbal task and the rank order of scores in the visual task. For each individual fracture type the number of correct responses on the verbal items was tested against the number of correct responses on the visual counterpart with the McNemar test. In all tests, a p-value < 0.05 was considered as the level of significance. RESULTS VISUAL TASK VERBAL TASK N of subjects N of subjects Total scores Total scores Figure 1. Frequency distributions of correct answers in the visual task (left diagram), and affirmative responses ( yes ) in the verbal recognition task (right diagram). The median score per participant was 3 (range 0-10) in the visual task (left diagram), and 6 (range 0-10) in the verbal task (right diagram). Legend: On the X-axis the total score of a subject is displayed. On the Y-axis the number of subjects with a particular total score is displayed. The frequency distributions of the total scores per participant are shown in Figure 1. It can be seen that the median was at the score of 3 on the visual recognition task, whereas the median for the total score on the verbal recognition task was 6.

51 50 Although, the overall performance differed between the two types of tasks, the rankorder of the total scores per participant on the verbal task showed a very strong association with the rankorder of the total scores on the visual recognition task (Spearman rank correlation coefficient (Rs) = 0.91). Figure 2 shows, that the mean score of verbal recognition (68% yes ) was statistically significantly greater than the mean score of visual recognition (33% correct). Turning to the performance per fracture type, the highest scores on both verbal and visual recognition were observed in respectively: Colles, Smith s, and distal forearm fracture. The percentage of correct responses per fracture type was also tested against our criterion of more than 80% correct responses. Although, 8 verbal tasks did not differ statistically significantly on the 80% criterion (Z-test), only 3 corresponding items (Colles, Smith s, and distal forearm fracture) satisfied the criterion in the visual task. According to the McNemar test, visual and verbal recognition were in accordance in 6 fracture types. However, a statistically significant difference was found in dorsal Barton s, volar Barton s, pilon, and radius & scaphoid fracture. Taken together, it may be concluded that in the group of 30 residents only Colles, Smith s, and distal forearm fracture met the 80% perfomance criterion. Almost no correct radiographic asssessment was found in 5 particular fracture types (pilon fracture, radial styloid process, volar and dorsal Barton s fracture, and radius & scaphoid fractures). In addition, the lowest performance on the corresponding verbal recognition task was found in chauffeur s and lunate load fracture. Apparently, specific intra-articular and fracture-dislocations were the difficult fracture types to recognize. Finally, an inventory was made of the incorrect descriptions on the 10 visual items for all 30 residents. It was found that the total of 68% (n=202) of responses was incorrect. A division could be made in either an incorrect diagnostic label, or an incomplete description. Remarkably, the labels Colles or Smith s were used in 98 out of 202 incorrect diagnoses. Figure 2. (see next page) Survey of the results of self-assessment of verbal knowledge, and the assessment of wrist radiographs (visual recognition). The data clearly show an empirical division in a group with more than 80% verbal and visual recognition, and a group of 7 fracture types with less adequate diagnostic performance. In the gray areas the observed measurement did not differ statistically significantly (Z-test) from the 80% performance criterion. Legend: The

52 51 maximum score per fracture item is 30. Fracture type Visual recognition Verbal recognition McNemar n correct (%) per item n yes (%) per item Colles 25 (83) 30 (100) NS Distal forearm 25 (83) 25 (83) NS Smith s 23 (77) 30 (100) NS Radial styloid process 16 (53) 20 (66) NS Barton s dorsal 4 (13) 20 (66) p < 0.05 Barton s volar 2 (7) 20 (66) p < 0.05 Pilon 2 (7) 19 (63) p < 0.05 Radius & Scaphoid 1 (3) 22 (73) p < 0.05 Chauffeur s 0 9 (30) NS Lunate load 0 8 (27) NS

53 52 Total 98 (33) 203 (68) p < 0.05 DISCUSSION This study showed that many residents have difficulties in recognizing the specific types of distal radial fractures. It appeared, that only Colles, Smith s, and distal forearm fracture were adequately recognized. It also occurred that the Colles and Smith s labels were most often used in incorrect visual assessments. In other words, the participants tend to label complex visual pictures in fracture types they already know. Most participants seemed to reduce the actual number of fracture types to 3 items in order to cope with 10 different distal radial fractures (6). Most likely, many subjects rely on a base-line knowledge which represents 3 fracture types. To our knowledge this observation has not been reported before, although it may be obvious to those involved in the management of wrist fractures and the training of residents. However, the results confirm the idea that a simple eponymous or synonymous classification system may not be the best strategy to deal with all distal radial fractures (2). A discrepancy between verbal and visual recognition was found in all fracture types. The observed difference may be explained by two reasons. Firstly, it appeared that visual and verbal recognition were statistically significant different in respectively, dorsal Barton s, volar Barton s, pilon, and radius & scaphoid fracture. The high scores on verbal recognition in these four items may reflect a tendency to overestimate the skill to visually recognize a fracture type on the X-ray. Overestimation in inexperienced physicians has been reported before and is also known as overconfidence bias (3). Verbal performance was significantly greater than visual performance in the mentioned four fracture types, and it seemed that overconfidence bias influenced diagnostic performance. Secondly, an explanation for the observed visual recognition rate of 12% in 7 fracture types may be a lack of verbal knowledge about the relevant distinctive features of distal radial fractures (3,6). The relevant distinctive features may not be known by inexperienced

54 53 physicians, because definitions and descriptions are often not described explicitely in publications. It seems plausible, that recognition might be improved if relevant distinctive features of distal radial fractures are described more systematically in handbooks or at educational courses. The drawback of simple classification systems with synonyms and eponyms, is that it is historical, not universal, and not treatment-based (7). The AO/ASIF s system for classification of fractures has been designed to overcome the above mentioned shortcomings, and may be a worthwhile tool for inexperienced physicians (5). The precise clinical value and the rate of agreement of the AO system for distal radial fractures has yet to be assessed. But, it has also been advocated by Mueller that the reliability of the AO system can only be improved if the relevant distinctive features of each fracture are emphasized even more better (5). REFERENCES 1. Cautilli RA, Joyce MF, Gordon E, Juarez R. Classifications of fractures of the distal radius. Clin Orthop and Rel Res 1974; 103: Fernandez DL. Classification. In: Fractures of the distal radius (Eds Fernandez D L and Jupiter J B), Springer-Verlag New York 1996: Gordon MJ. A review of the validity and accuracy of self-assessments in health professions training. Academic Medicine 1991; 66: Melone CP. Distal radius fractures: Patterns of articular fragmentation. Orthop Clin North Am 1993; 24: Mueller ME. Two pamphlets of the Comprehensive Classification of Fractures, edition AO/ASIF Documentation Center Davos, Switzerland Norman GR, Coblentz C L, Brooks L R, Babcook CJ. Expertise in visual diagnosis: A review of the literature. Academic Medicine 1992; 67 Oct suppl: s79-s Oskam J, Kingma J, Venekamp R, Klasen H J. Colles and Smith as scholars for residents. Nederlands Tijdschrift voor Traumatologie 1994; 2: Oskam J, Graaf de JS, Klasen HJ. Fractures of the distal radius and scaphoid. J Hand Surg 1996; 21B: Sheridon CL. Fundamentals of experimental psychology. Holt, Reinhardt, and Winston New York Solgaard S. Classification of distal radius fractures. Acta Orthop Scan 1984; 56:

55 54 CHAPTER 5 THE BASIC CATEGORIES OF THE AO/ASIF S SYSTEM AS A FRAME OF REFERENCE FOR CLASSIFYING DISTAL RADIAL FRACTURES. J Oskam, J Kingma, H J Klasen. Department of Surgery, University Hospital Groningen Groningen, the Netherlands

56 Submitted 55

57 56 Since 1987, the AO/ASIF system for classification of fractures is being used to predict outcome and to assist in choosing an appropriate treatment regime (2). The AO system appears to be an attractive frame of reference, because the Orthopaedic Trauma Association in the United States of America has recently adopted the AO system as the standard of fracture classification. Moreover, the AO classification is quite often teached in postgraduate courses and handbooks. Although the AO system is used worldwide, only few studies to interrater agreement or reliability have been published as far as we know. In respect to distal radial fractures, only one study has been published which reported specifically on agreement of the AO classification (1). In this study, the focus of interest was the interrater agreement in 36 assessors, with varying clinical experience, in a selected sample of 30 fractures. Good agreement was observed for experienced surgeons only (kappa value = 0.68). In addition, it appeared that agreement was just good for the level of the 3 basic types (A, B, and C), while the kappa values for the lower levels of the 9 main groups and the 27 subgroups were 0.48 (moderate agreement) and 0.33 (fair agreement), respectively. May be, it is not surprising that the rate of agreement diminished with ongoing refinement of the classification system into subclasses. However, the usefulness of a classification method depends largely on the reliability of the system. Agreement in the basic categories should be almost perfect before classification at lower levels is performed, because further refinement in subclasses will inevitably result in a decrease of agreement. Therefore, it may be interesting to analyze which problems may rise in the allocation of fracture types into the 3 basic classes. And, how some issues can be resolved before other levels of classification are studied. The purpose of the present study was to investigate how the AO methodology may be employed to assign distal radial fractures which are usually seen in the emergency unit, and to determine the effect of consensus stimulation by the so-called Delphi approach on observer agreement. A second concern was to describe the underlying causes of possible disagreement between two observers.

58 57 MATERIAL AND METHODS Study design A random sample of 124 cases was drawn out of 385 patients older than 18 years (median age 58 years), who visited the accident and emergency department in 1994 for a distal radial fracture. A fresh fracture and clear AP and lateral wrist X- rays were available for all 124 patients. The X-rays were presented in a random order, and were assessed by two expert surgeons who treat wrist fractures regularly. The two observers assessed the X- rays independently. No information was given about the performed treatment and outcome. The observer was asked to classify the fracture according to the AO/ASIF s classification system on a separate sheet. The AO/ASIF classification for distal radial fractures (no. 23.x-x) is composed of 3 basic fracture types, 9 main groups, and 27 subgroups (2). The 3 basic fracture types are extra-articular fractures (class A), partial intra-articular fractures (class B), and complete intraarticular fractures (class C). The observers were asked to assign the fractures to one of the basic groups A, B, or C, and to group D if a fracture could not be attributed to a particular AO category (2). The observers were allowed to use the scheme of the classification system with examples as provided by the AO/ASIF, and there was no limitation of time to complete the test. After the test was completed the observers were invited to join a consensus meeting. The Delphi approach was used to reach a consensus regarding the most appropriate assignment to the 3 basic AO categories for which a disagreement between the two observers was found (3). The steps involved in this process are summarized below. As a first step in developing a consensus, the two observers employed the criteria and the assumptions of the AO classification for the 3 basic classes. Having agreed on these assumptions independent assignment of the fractures to either class A,B,C or D was then made again by the two observers. During the Delphi approach conflicting arguments in those fractures with disagreement were discussed between the two observers, and the X-rays were viewed again. The

59 58 results of the independent ratings were compared and discrepancies identified. Then, a discussion followed on the basis of relevant distinctive features of distal radial fractures (5). Most discrepancies were resolved by discussion of the relevant distinctive features by both observers and an independent chair. Finally, it was decided whether consensus was reached. The assignments for which disagreement remained were identified and recorded to allow further analysis. Statistics For each X-ray, a pair wise comparison of the answers of the two observers was performed in a cross-table. The percentage of agreement was computed both for before and after the Delphi meeting. Furthermore, the fracture types in which disagreement occurred were located in the cross-table. Cohen s kappa statistics was used to adjust for overall agreement attributable to chance (6). For the interpretation of the degree of agreement the following criteria were used: Kappa values 0.0 through 0.20 represent slight agreement, 0.21 to 0.40 fair agreement, 0.41 to 0.60 moderate agreement, 0.61 to 0.80 good agreement, above 0.81 is considered almost perfect agreement (4). Categorization By Observer 2 A B C D A Categorization By Observer 1 B C D Table 1a. Cross tabulation of the assessments (n=124) of observer 1 and observer 2 according to the four basic categories (A,B,C,D) of the AO/ASIF s classification before the Delphi approach was undertaken. The figures represent absolute numbers. Agreement for the categories A,B,C, and D

60 59 are shown in bold numbers. Disagreement between the observers can be seen in not-bold printed numbers. The sum of the bold numbers (n=99) is the total agreeement in all 124 fractures. RESULTS The overall interrater agreement was good because Cohen s kappa value was Table 1a shows the cross tabulation of the assignment of 124 fractures to the basic categories by the two observers. If both observers agree completely all observations would lie on the main diagonal. So, each element on the diagonal line represents the degree of agreement for a particular category. It can be seen that the 2 observers agreed on 62 assignments to class A, i.e. 50% (see also Table 1b) of the 124 X-rays. The majority (72%) of the fractures was assigned to respectively category A (50%) and category C (22%). The sum of the elements on the diagonal line represents the overall agreement between the 2 observers, e.g. agreement was found for 99 fractures (Table 1a), or 80% (Table 1b) across all categories. Categorization By Observer 2 A B C D A Categorization By Observer 1 B C D Table 1b. Cross tabulation of the assessments (n=124) of observer 1 and observer 2 according to the four basic categories (A,B,C,D) of the AO/ASIF s classification before the Delphi approach was

61 60 undertaken. The figures represent percentages. The percentage of agreement for the categories A,B,C, and D are shown in bold numbers. Disagreement between the observers can be seen in notbold printed numbers. The sum of the bold numbers (80%) is the total percentage of agreeement in all 124 fractures. Statistically, Cohen s kappa value for overall agreement was The elements beside the diagonal line represent the degree of disagreement between the two observers. Table 1 shows that the highest degree of disagreement occurred for the categories A and C. Observer 1 assigned 5 fractures to category A, whereas Observer 2 assigned the same fractures to category C (Table 1a). On the other hand, Observer 1 assigned 7 fractures to category C, whereas Observer 2 assigned the same fractures to category A. Thus, analysis of Table 1 shows an A versus C reversal shift, i.e. a controversy between the assignments to either the A or C category. Categorization By Observer 2 A B C D A 56 2 Categorization By Observer 1 B 7 C 5 23 D 1 6 Table 2. Cross tabulation of the assessments (n=124) of observer 1 and 2 according to the four basic categories (A,B,C,D) of the AO/ASIF s classification after the Delphi approach was undertaken. The figures represent percentages. The percentage of agreement for the categories A,B,C, and D are shown in bold numbers. Disagreement between the observers can be seen in the not-bold printed numbers. The sum of the bold numbers (92%) is the total percentage of agreeement in all 124 fractures. Statistically, Cohen s kappa value for overall agreement was 0.86.

62 61 The results of the Delphi meeting can be seen in Table 2. It appeared that the overall agreement was 92%, which was 12% higher than the observed 80% before the meeting (Table 1b). Due to the consensus meeting the kappa value for the 3 basic classes increased from 0.65 to 0.86 (excellent agreement). Agreement on category A increased from 50% to 56%. It appeared that about 5% of this increase came from category D fractures. Another remarkable shift can be observed in category B. In this group it was found that agreement rose from 3% to 7%. During the meeting it occurred that disagreement on a B versus C fracture, or a B versus A fracture could be solved. Despite the increase of agreement in the A and B fractures, a part of disagreement still remained: some controversy between class A and C fractures ( reversal shift ) was still observed, although the A/C reversal shift decreased from 10% to 7% of cases after the meeting. With respect to the D fractures, disagreement was found in 7% of cases before the meeting, and decreased to 1% thereafter. The reason for this shift was that in these particular cases an undisplaced fracture was considered to be displaced by both observers. These fractures were assigned as category A. However, after the meeting 8 fractures were still considered to be undisplaced and remained classified as category D. DISCUSSION The AO system showed to be a useful tool for classifying distal radial fractures, since good agreement (kappa value 0.65) was observed before the Delphi meeting. The present finding that the presence or absence of articular involvement can be assessed consistently if classification is undertaken by experienced observers has also been reported by Kreder et al. However, since the purpose of their study was just to quantify agreement, analysis to specific causes of disagreement was not undertaken. Presently, a qualitative analysis to the underlying mechanisms of disagreement was performed by employing the Delphi approach. It was found that non-displaced fractures were a major source of controversy. These results are in agreement with Kreder s findings, and it seems

63 62 rational that a separate class for undisplaced fractures has to be defined in the AO/ASIF s system to resolve this important issue (1). Although an increase in interrater agreement was found, perfect agreement on the 3 basic classes could not be reached. May be, the observed A/C reversal shift is one of the causes of disagreement. Apparently, the controversy between extra- (A) and intra-articular (C) fractures is a basic problem, because disagreement remained in 7% of cases even after extensive discussion. It might be hypothesized that the main cause of disagreement is a shortage of information due to the radiographic imaging technique. In other words, technological shortcomings may bias classification reasoning with respect to intra-articular involvement. Therefore, radiological innovations or additional information, for instance by routinely employing oblique directions of the distal radius, may increase agreement on articular involvement in the distal radius. Many conflicting assessments could be resolved during the consensus meeting by discussing the particular X-rays. It appeared that the sources of conflict were differences in opinion about relevant distinctive features. However, on the basis of these relevant distinctive features, the observers were willing to agree on the choice of a particular classification group. But disagreement remained in several cases even in experienced observers. It may be impossible to reach perfect agreement due the above-mentioned A/C reversal shift, or because fracture types are not good enough defined. However, we experienced that discussion on the basis of relevant distinctive features was a good starting point to improve agreement. It appeared that agreement was only good for the level of the 3 basic classes. Most likely, the clinical relevance of using only 3 classes is too limited to assess all types of distal radial fractures. However, we feel that the observed problems at the basic level, like the issues of the undisplaced fractures and articular involvement, have to be resolved first. Otherwise, this noise will continue to create problems at lower, clinically more meaningful levels in the classification system. We feel that radiological innovations are not the only measures to improve agreement on the AO/ASIF s classification system for distal radial fractures. But intercollegiate

64 63 discussion on relevant distinctive features may also be important to reach the goal of good agreement. REFERENCES 1. Kreder HJ, Hanel DP, McKee M, Jupiter J, McGillivary G, Swiontkowski MF. Consistency of the AO fracture classification for the distal radius. J Bone Joint Surg 1996; 78B: Mueller ME, Nazarian S, Koch P, Schatzker J. The comprehensive classification of fractures of long bones. Springer-Verlag Berlin, Germany McKenzey EJ, Steinwachs DM, Shankan BS. An ICD-9-CM to AIS conversion table development and application. Proc Am Assoc for Auto Med 1986; 30: Landis JR, Kochj GG. The measurement of observer agreement for categorial data Biometrics 1977; 33: Oskam J, Kingma J, Karthaus AJM, Klasen H J. Recognition of 10 distal radial fractures types by residents submitted. 6. Reynolds H. The analysis of cross-classifications. Mcmillan Publishing Co. New York, USA 1977.

65 64 CHAPTER 6 FRACTURES OF THE DISTAL RADIUS AND SCAPHOID J Oskam, J S de Graaf, H J Klasen. Department of Surgery, University Hospital Groningen Groningen, the Netherlands Journal of Hand Surgery, 1996; 21B:

66 65 The usual conservative treatment for isolated distal radial fractures is to apply ligamentotaxis across the radiocarpal joint by traction with immobilization of the wrist in a neutral position. However, it is generally believed that tractional forces should be avoided in scaphoid fractures, while the wrist is best positioned in radial deviation with immobilization of the thumb. In the case of a combination of fractures of the distal radius and scaphoid, each immobilization technique may have opposite effects on fracture healing. Therefore, one might expect either an increased rate of scaphoid nonunion or malunion of the distal radius. We have treated 23 patients with simultaneous fractures of the distal radius and scaphoid over a period of 14 years. A below elbow cast including the thumb was used as standard treatment, because we feared more scaphoid than distal radius complications. In the present study, we have evaluated the results of our treatment policy. PATIENTS AND METHODS During the period 1980 to 1993, a fracture of the distal radius and the scaphoid was diagnosed in 23 patients. There were 10 women and 13 men, with a median age of 39 years (range 18-74). The dominant hand was involved in 9 patients. The injury resulted from an accidental fall in 17, sports in 3 and traffic accidents in 3 cases. X-ray assessment The scaphoid and distal radial fracture were classified separately. The scaphoid fracture was assessed for site and displacement. Displacement of the scaphoid fracture was defined as displacement of at least 1 mm. The scaphoid was divided in 3 parts to describe the fracture localization (2). The distal radial fracture was classified by type, and involvement of the radiocarpal joint surface was also assessed (3). During follow up, X-rays of the distal radius and scaphoid were used to investigate signs of disturbed bone healing. To assess malunion of the distal radius, the radiocarpal angle and radial shortening were measured (7). The scaphoid was

67 66 assessed for pseudarthrosis, avascular necrosis, or cysts. Finally, signs of posttraumatic osteoarthritis and carpal instability were noted. Treatment A below elbow cast including the thumb with the wrist immobilized in radial deviation and flexion was applied in 18 patients treated conservatively. The mean immobilization period was 9 weeks, and depended on healing of the scaphoid bone. Closed reduction of the distal radial fracture was performed in 9 of 18 conservatively treated wrists. Primary operative treatment was undertaken in 3 patients. A secondary operation was performed in 2 patients in whom redisplacement of the distal radius occurred during conservative treatment. An unstable distal radial fracture was the indication for surgery in all cases. Screw fixation of the radius was performed in a palmar Barton s fracture. An external fixator was applied in 2 other cases with a comminuted intra-articular distal radial fracture. Closed re-reduction and K-wire pinning was performed in the 2 patients with radial redislocation. Concomitant fixation of the scaphoid fracture with a cannulated cancellous bone screw was performed in 4 patients. Follow-up The follow-up study comprised both clinical and X-ray examination. Healing of the fractures was radiologically monitored in all patients a 5 day, 2, 6, 9, and 12 weeks intervals. In total, 21 patients were eligible for examination, because 2 patients were deceased at the time of study. Pain, range of wrist motion, and grip strength were assessed. The functional end results were judged as good when no pain occurred with use, and fair if pain and moderate discomfort was present with use, provided that the patients had completely returned to normal activities. The functional end result was considered to be poor in all other circumstances. RESULTS The mean follow up period was 7 years (range 1-13). The most serious complication of fracture healing was redisplacement in 3 of 9 initially dorsally-

68 67 displaced distal radial fractures. An extra-articular Colles type fracture with severe comminution of the dorsal cortex was present in these 3 patients. Closed reduction and trans-styloid Kirschner-wire fixation was performed to prevent malunion of the distal radius in these patients. Post-traumatic carpal instability was not observed in any wrist. Healing of the scaphoid fracture was uncomplicated in all 23 wrists. It appeared that all serious complications occurred on the side of the distal radial fracture. Fracture type It can be seen in Table 1, that all scaphoid fractures were localized in the middle or distal third. Displacement greater than 1 mm of the scaphoid fracture existed in 4 of 23 patients. Rotational subluxation or carpal instability was not present. An extra-articular fracture of the distal radius was found in 15 patients, and in 7 dorsal displacement (Colles fracture) was observed. An intra-articular distal radial fracture was observed in 8 of 23 wrists. Distal radius Scaphoid Middle third Waist Distal third Extra-articular Undisplaced Colles Intra-articular 2 4 Chauffeur s 1 Barton 1 Table 1. Radiological classification of 23 simultaneous fractures of the distal radius and the scaphoid. Wrist function Overall, 21 patients were satisfied with the functional end result and had resumed normal daily activities. Operative treatment had been undertaken in 4 of them.

69 68 Wrist pain was reported by 5 patients. Diminished dorsal flexion with disturbed forearm rotation was observed in 4 patients, of whom 3 were treated surgically. Subjective loss of grip strength was found in 2 patients with fair and poor functional results. After healing of the distal radius, shortening (2-7 mm) was observed in 6 wrists. Diminished wrist motion was found in 4 patients. Wrist pain and serious limitation in daily life was found in only 2 patients. One patient could not resume his job and complained of pain with severe loss of wrist motion due to radiocarpal osteoarthritis (a poor result), while another patient suffered from malunion of the distal radius. A good functional result was eventually observed in 18 patients. DISCUSSION This study confirms that in most simultaneous fractures of the distal radius and scaphoid good results can be obtained with conservative treatment (4,6). However, the finding that nearly all problems with fracture healing occurred in the distal radius has not been reported before. The redisplacement of three distal radial fractures might well have been caused by the position of the wrist in the below elbow cast. In this series the wrist was immobilized in radial deviation, a position which can provoke radial displacement because it allows the brachioradial muscle to act on the distal radius. Although we cannot be certain, we think that radial deviation of the wrist might have contributed to redisplacement of the distal radius. Primary surgery was necessary in a minority of cases. The indication for surgery in the three primarily operated patients was an unstable, intra-articular distal radial fracture, in which the risk of malunion with conservative treatment was thought to be unacceptable. We would also have operated on these wrists had the fracture been isolated, so the presence of the scaphoid fracture did not influence the decision. In our opinion, standard indications for operation cannot be given, and surgical treatment should be tailored to the individual patient. Generally, the decision whether to operate on the simultaneous fractures should be based on the

70 69 same criteria as isolated fractures, e.g. an unstable, displaced scaphoid fracture (2,5), an unstable, displaced distal radial fracture (3), and carpal instability. All scaphoid fractures healed normally, and avascular necrosis was not observed. The explanation for this observation is most likely the fact that 19 out of 23 scaphoid fractures were undisplaced and localized to the waist. It has been described previously that healing of fractures near or at the waist of the scaphoid is complicated in only 5% of cases (2). However, we initially thought that the scaphoid fracture would produce the more serious complications, so a below elbow cast including the thumb instead of a dorsal splint was used. It appears that the outcome of the combined fractures is determined more by the distal radial fracture. It has been reported that the type of cast used to immobilize stable scaphoid fractures does not affect the incidence of non-union or other complications of fracture healing (1), and that good results could be achieved by applying a dorsal splint with the wrist in neutral position. Immobilization of the thumb is not necessary and reduction of the scaphoid fracture was not lost with tractional forces across the wrist. Consequently, there is strong evidence that treating a scaphoid fracture with the wrist in neutral position is not detrimential. As a result of this study we shall apply a dorsal splint with the wrist in neutral position for 6 weeks for non-operative treatment in future cases, because there is more likelihood of redisplacement of the distal radial fracture than non-union of the scaphoid. REFERENCES 1. Clay NR, Dias JJ, Costigan PS, Gregg PJ, Barton NJ. Need the thumb be immobilised in scaphoid fractures? A randomised prospective trial. Journal of Bone and Joint Surgery 1991; 73B: Cooney WP, Dobyns JH, Linscheid RL. Fractures of the scaphoid: a rational approach to mangement. Clinical Orthopaedics and Related Research 1980; 149: Jupiter JB. Current concepts. Review fractures of the distal end of the radius. Journal of Bone and Joint Surgery 1991; 73A: Smith JT, Keeve JP, Bertin KC, Mann RJ. Simultaneous fractures of the distal radius and scaphoid. Journal of Trauma 1988; 28:

71 70 5. Szabo RM, Manske D. Displaced fractures of the scaphoid. Clinical Orthopaedics and Related Research 1988; 230: Tountas AA, Wadell JP. Simultaneous fractures of the distal radius and scaphoid. Journal of Orthopaedic Trauma 1988; 1: Warwick D, Prothero D, Field J, Bannister G. Radiological measurement of radial shortening in Colles fracture. Journal of Hand Surgery 1993; 18B: 50-2.

72 71

73 72 CHAPTER 7 DORSAL FRACTURE-DISLOCATION OF THE RADIOCARPAL JOINT: A FOCUS ON OPERATIVE TECHNIQUE. J.Oskam, RAEC Hermens, HJ Klasen. Department of Surgery, University Hospital Groningen Groningen, the Netherlands Submitted

74 73 Dorsal radiocarpal fracture-dislocation is defined by dorsal carpal displacement caused by disruption of the radiocarpal ligaments, while the associated distal radius fracture has to be confined to the rims and the styloid process (3). However, a combination of radiocarpal and ulnocarpal ligamentous rupture, and fracture of the ulnar styloid process are usually also present. Several cases have been reported in the past 25 years (2, 4, 6, 8). In these papers, mainly clinical and taxonomic aspects have been discussed. However, specific methods of operative treatment to improve functional outcome have not yet been reported. In this paper we will report on 6 patients which have been operated at the University Hospital Groningen in the period During this period we performed 3 different operation techniques. We will discuss our experiences with operative management, furthermore the functional results will be assessed. CASE REPORTS Case 1. A 40 years old man sustained multiple injuries while demolitioning a wall of a house. A part of the wall fell on the patient, as he tried to hold back the wall with his left wrist dorsiflexed. An open dorsal fracture-dislocation of the left wrist resulted with a large transverse wound volarly. Besides the fracture-dislocation, a dorsal intercalated segment instability (DISI) with ulnar carpal translocation was also present. Sensation was impaired in the second, third and fourth finger. Other injuries were a large knee wound, and a large degloving injury at the lumbar region of the back. At operation, debridement was performed and flexor tendons, arteries, and median nerve were found to be intact. The volar carpal ligament was ruptured. Following reduction, the styloid process of the radius was fixed with 2 K-wires. A plaster cast was applied to maintain carpal reduction. After 6 weeks the cast and the K-wires were removed, however, the radial styloid process was redislocated. Three years later, the patient had returned to his previous work without complaining of pain and with intact rotation of the forearm. However, severe osteoarthritis in the carpus was present and wrist movement was severely impaired resulting in a functional arthrodesis. Case 2. A 44 years old municipal worker fell from a 9 meter height and sustained fractures of the sternum, the first lumbar vertebra, and a dislocation of the left wrist. The skin, circulation and sensation of the hand were intact. Roentgenograms showed a dorsal dislocation of the carpus with a comminuted fracture of the styloid proces and the dorsal rim of the distal radius. Closed reduction and fixation with a K-wire was performed, and the wrist was immobilized with a dorsal plaster splint.

75 74 The K-wire was inserted proximally, and dorsally over the radiocarpal joint. After the procedure the carpus seemed to be stable. An additional release of the volar carpal ligament was also performed. Unfortunately, during the immobilisation period, due to loosening of the K-wire dorsal subluxation of the carpus occurred. Nine years after the injury, the patient complained of pain while working. Physical examination showed 40% impairment of flexion and extension compared to the opposite wrist, with intact forearm rotation. Roentgenograms showed severe osteoarthritis of the radiocarpal joint. The patient needed to wear a wrist orthesis to be able to work. Case 3. A 37 years old construction worker fell from a 6 meter height. He sustained head wounds and a fracture of the right radial styloid process. The left wrist showed a closed dorsal radiocarpal dislocation without impairment of circulation and sensation. Roentgenograms showed a radiocarpal dislocation with associated fractures of the radial and ulnar styloid process, as well as, a fracture of the dorsal radial rim. Closed reduction and K-wire fixation was performed, as well as release of the volar carpal ligament and application of a dorsal plaster splint. After inserting a K-wire on the wrist dorsally, radiocarpal stability could only be achieved after fixation of the radial styloid process with a second K-wire. Although carpal reduction could be maintained, redislocation of the styloid process occurred during the first weeks, postoperatively. Five years after the injury the patient had not returned to his previous job. He complained of wrist pain, and a loss of 50% of wrist movement with intact forearm rotation was present. The roentgenograms showed radiocarpal osteoarthritis of moderate degree. Case 4. A 31 years old man fell while racing with a motorcycle. A fracture-dislocation of the right humerus and a dislocation of the left wrist resulted. Skin, circulation and sensation of the left hand were intact. Roentgenograms showed a dorsal dislocation of the carpus with concomitant fractures of the radial and ulnar styloid processes and of the dorsal rim of the radius. During operation, closed reduction and trans-articular external fixation was performed. Furthermore, release of the volar carpal ligament was carried out. Radiocarpal reduction was maintained with an external fixator, which was removed after 5 weeks. It was not possible to fixate the radial styloid process because of fracture comminution. Postoperatively, no signs of redislocation were observed. Three years later, the patient had returned to his previous job and did not complain of wrist pain. Physical examination showed a limited wrist movement of 40% due to moderate osteoarthritis of the radiocarpal joint. Case 5. A 18 years old man sustained a car accident. A dorsal fracture-dislocation of the right wrist resulted. Skin, circulation and sensation were intact. A CT-scan showed that a DISI, Scaphoid- Lunate (SL) dissociation and ulnar carpal translation was also present. At operation, open reduction and screw fixation of the radial styloid process and the dorsal rim was performed. Radiocarpal alignment was maintained with a protruding dorsal T-plate, which was fixated dorsally on the distal

76 75 radius. Because of ulnocarpal instability the ulnar styloid process was also fixated with a 2.7 mm screw. Practising was started after 2 weeks of immobilization. Removal of the plate was performed 8 weeks, postoperatively. After 1 year, wrist movement was limited to 40% and pain was present. Severe osteoarthritis of the radiocarpal joint and proximal carpal row existed. Carpal instability with scaphoid-lunatum (SL)-dissociation and ulnar translation was still present. Case 6. A 22 years old farmer sustained multiple injuries in a car accident. These comprised brain concussion, a halfsided maxilla fracture, and a dorsal fracture-dislocation of the left wrist (See Figure 1a). The dislocation was reduced, and the volar carpal ligament was released. After closed reduction, the radial styloid process was fixated with a cancellous bone lag screw through a separate small radial incision. A second dorsal longitudinal incision between the second and third extensor compartment was also made to perform open reduction and stabilisation of the carpus. Good radiocarpal alignment was maintained with a protruding T-plate which was fixed on the distal radius (See Figure 1b). The plaster cast was removed after two weeks, and practising was started. Following removal of the plate at 6 weeks, full return to normal activities occurred. After three years, the patient experienced no wrist pain, although wrist movement was limited to 50%, probably due to moderate osteoarthritis (See Figure 1c). DISCUSSION The histories of the six patients show that dorsal radiocarpal fracture-dislocation usually occurs in multiple injured victims. Most likely, extreme dorsiflexion of the wrist with pronation of the forearm on the fixed hand is the injury mechanism. Apart from fracture-dislocation, median nerve palsy and soft tissue laceration are likely to develop. In some instances, the volar carpal ligament is also ruptured, but if not, release is strongly indicated (6). Remarkably, circulation of the hand commonly is not compromised. One of the most striking clinical features is a severely unstable radiocarpal joint, for which operative treatment is usually mandatory (2). It has been advocated, that closed reduction and K-wire fixation followed by immobilisation with a splint may create adequate stability (4, 5). We followed this policy in three cases. Surprisingly, a complication occurred in all three patients. Either carpal subluxation or redislocation of the radial styloid process was observed. Most likely, forces on the radiocarpal joint could not be compensated by

77 76 K-wires and a dorsal splint. It was also observed that radiocarpal stability could only be reached if the radial styloid process was fixated. Therefore, we changed our policy and fixated the radial styloid process with a lag screw in Cases no. 5 and 6. As a result, carpal reduction was maintained and redislocation of the radial Figure 1. a Roentgenograms of Case 6, showing a dislocation of the entire carpus dorsally with a fracture of the dorsal rim and styloid process of the distal radius. b The same wrist, postoperatively. Fixation of the radial styloid process with a screw, and stabilization of the carpus with a dorsal, protruding T-plate is performed. c The radiological result three years after operation. Moderate osteoarthritis and calcifications around the radiocarpal joint are present.