Apparently isolated partial articular fractures of the radial head: Prevalence and reliability of radiographically diagnosed displacement

Apparently isolated partial articular fractures of the radial head: Prevalence and reliability of radiographically diagnosed displacement

Apparently isolated partial articular fractures of the radial head: Prevalence and reliability of radiographically diagnosed displacement Job Doornber...

395KB Sizes 0 Downloads 23 Views

Apparently isolated partial articular fractures of the radial head: Prevalence and reliability of radiographically diagnosed displacement Job Doornberg, MS,a,b Andreas Elsner, MD,a Peter Kloen, MD, PhD,c René K. Marti, MD, PhD,c,d C. Niek van Dijk, MD, PhD,c and David Ring, MD,a,f Boston, MA, and Amsterdam, The Netherlands

The purpose of this study was to measure the prevalence and reliability of the radiographic diagnosis of displacement of apparently isolated partial articular radial head fractures and use these factors to assess treatment considerations. Among 119 radiographically visible partial fractures of the radial head not associated with other wrist, forearm, or elbow injury, 101 were classified as Mason type 1 (85%), 11 as borderline between Mason type 1 and Mason type 2 fractures (9%), and 7 as Mason type 2 fractures (6%) according to Broberg and Morrey’s modification of the Mason classification. The intraobserver reliability of the classification of Mason type 1 and type 2 fractures was excellent (mean ␬, 0.85), but the interobserver reliability was only moderate (multirater ␬, 0.45). Because apparently isolated, stable partial fractures of the radial head are infrequently displaced and observers have moderate disagreement regarding the diagnosis of displacement, it is likely that displacement is overdiagnosed. (J Shoulder Elbow Surg 2007;16:603-608.)

M ason’s distinction of nondisplaced fractures (type 1), displaced fractures of part of the radial head (type 2), and displaced fractures of the entire head (type 3) remains the most widely used system for classifying fractures of the radial head. Mason16 did not quantify or define displacement. The suggestion of Broberg and Morrey6 that a partial radial head fracture must be of sufficient size (at least 30% of the articular surface) and displacement (at least 2 mm) to be

From the aHand and Upper Extremity Service, Department of Orthopedic Surgery, Massachusetts General Hospital, and fDepartment of Orthopedic Surgery, Harvard Medical School, Boston, and bOrthopedic Research Center, cDepartment of Orthopaedic Surgery, Academic Medical Center, and dOrthotrauma Research Center, University of Amsterdam. Reprint requests: David Ring, MD, Massachusetts General Hospital, Yawkey Center, Suite 2100, 55 Fruit St, Boston, MA 02116. (E-mail: [email protected]). Copyright © 2007 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/2007/$32.00 doi:10.1016/j.jse.2006.10.015

considered a displaced fracture (Mason type 2, as opposed to nondisplaced Mason type 1) is widely used in treatment decisions. This is referred to as a modified Mason classification herein. Although some investigators have questioned the reliability of these distinctions17 and the recommendation is not very well supported by data,8,12,13,18,22 apparently isolated partial radial head fractures that meet these criteria are generally considered candidates for surgery. Nonetheless, Herbertsson et al10 recently reported good long-term results with nonoperative treatment of displaced, uncomplicated modified Mason type 2 fractures. It has been noted that displacement of a fracture of the radial head is likely to indicate associated ligament injury of the elbow or forearm.7 Displaced fractures of the radial head with no other fractures or ligament injuries may be relatively uncommon. In general, the relative infrequency of a condition magnifies the shortcomings of a test used for diagnosis of that condition,3,4,14 with a resulting tendency to overdiagnose the condition. If radiographic interpretations of displacement are somewhat unreliable and displacement is uncommon, then radiographic diagnostic methods may tend to overdiagnose displacement (high false-positive rate). If one considers this tendency to overdiagnose displacement in context with the good long-term results of nonoperative treatment for apparently isolated displaced fractures, it can be argued that surgeons should err toward nonoperative treatment of apparently isolated partial fractures of the radial head. These issues may be particularly important where medical resources are limited. The purpose of this study is to report the prevalence of apparently isolated modified Mason II fractures and test the intraobserver and interobserver reliability of the criteria for considering operative treatment, namely displacement, according to the criteria of Broberg and Morrey.6 We evaluated the prevalence of apparently isolated displaced partial radial head fractures, according to the criteria of Broberg and Morrey,6 presenting to a single surgeon to answer our first study question:

603

604

Doornberg et al

How common is radiographic diagnosis of apparently isolated displaced (Mason type 2) partial radial head fractures according to their criteria? In addition, we tested the intraobserver and interobserver reliability of radiographic criteria for displacement of apparently isolated partial articular fractures of the radial head to find an answer to the following question: How reliable are the criteria for operation—that is, greater than 2 mm of displacement and at least 30% of the articular surface—according to Broberg and Morrey? We hypothesized that displaced partial radial head fractures are relatively uncommon and unreliably identified. MATERIALS AND METHODS We retrospectively reviewed the billing records of a single surgeon. Between 2001 and 2004, 286 radial head fractures were treated. Exclusion criteria included fractures of the radial neck (nonarticular fractures), fractures associated with other bone or ligament injuries in the elbow or forearm, fractures of the entire head (Mason type 3), and fractures diagnosed based on clinical factors and radiographic signs of an elbow effusion (so-called radiographically occult fractures). The radiographs of these 286 patients were evaluated by use of a protocol approved by our human research committee. All patients had anteroposterior, lateral, and oblique radiographs, but radial head– specific views were rarely available. The patients excluded had 22 radial neck fractures, 90 radial head fractures (all displaced) that were part of an elbow or forearm fracture-dislocation (42 of these were partial-head fractures), 4 apparently isolated Mason type 3 fractures, and 51 radiographically occult radial head fractures. Thus, the cohort of radiographically visible, apparently isolated partial fractures of the radial head that formed the basis of this study comprised 119 patients. There were no radiographic or clinical signs of associated fracture or ligament injury of the forearm or elbow in any patient.

Classification Together, 2 observers classified these 119 fractures according to the modification of Broberg and Morrey5 of the original classification of radial head fractures of Mason.16 According to the Broberg and Morrey system, a Mason type 1 fracture of the radial head is present when the fracture fragment is displaced by less than 2 mm or involves less than 30% of the articular surface, a Mason type 2 fractures is displaced by at least 2 mm and involves at least 30% of the articular surface, and a Mason type 3 fracture involves the entire head. When the 2 observers had difficulty agreeing on the degree of displacement, the fracture was considered to be on the borderline between Mason types 1 and 2 and was classified as such.

Intraobserver and interobserver reliability of classification of modified Mason type 2 fractures On the basis of the classifications of the 2 observers, 30 fractures were randomly selected for the evaluation of in-

J Shoulder Elbow Surg September/October 2007

traobserver and interobserver reliability, comprising 13 fractures that both observers agreed were modified Mason type 1, 7 fractures that both observers agreed were modified Mason type 2, and 10 fractures classified as borderline (between modified Mason types 1 and 2). This method of selecting cases for review was chosen because although it may tend to exaggerate disagreement, it most closely reflects clinical practice, as only notable fractures will be considered for surgery and will be closely scrutinized to determine whether they fit the Broberg and Morrey criteria for displacement. The radiographs of these 30 fractures were presented to 5 independent observers with different levels of experience and specialties: 1 specialized upper extremity surgeon, 1 upper extremity fellow, 1 orthopaedic traumatologist, and 2 general orthopaedic surgeons with interest in trauma care. The 30 cases were randomized and presented in 2 rounds separated by 2 weeks to evaluate intraobserver reliability. The observers received a written and illustrated description of Broberg and Morrey’s modification of the Mason classification. The criteria for defining displacement of partial-head fractures were highlighted, and understanding was confirmed by the study staff before the interpretations were begun. The observers were blinded to the purpose of the study, were not involved in the care of these patients, and were not aware of the treatment and outcomes of the individual patients.

Statistical analysis The ␬ value was used to measure the intraobserver and interobserver reliability of fracture classification. The ␬ value is a chance-corrected measure of agreement comparing the observed measure of agreement with the level of agreement expected by chance alone.15 It has been the most commonly used statistic to describe agreement in a variety of intraobserver and interobserver studies since its introduction by Cohen in 1961.2,9,19,20 Interobserver reliability among different observers was calculated by use of the multirater ␬ measure as described by Siegel and Castellan.21 The analysis of interobserver reliability was based on both the first and second rounds of observations. ␬ Values were assigned to subdivisions according to the report of Landis and Koch15: slight agreement, 0.00 to 0.20; fair agreement, 0.21 to 0.40; moderate agreement, 0.41 to 0.60; substantial agreement, 0.61 to 0.80; and almost perfect agreement, greater than 0.81. A ␬ value of 0 represents no agreement and 1.0 represents perfect agreement. Statistical analysis was performed with SPSS software (SPSS, Chicago, IL).

RESULTS Prevalence

According to the modified Mason classification by Broberg and Morrey,6 101 fractures were classified as modified Mason 1 (85%), 11 as borderline between type 1 and type 2 (9%), and 7 as modified Mason type 2 (6%).

J Shoulder Elbow Surg Volume 16, Number 5

Intraobserver and interobserver reliability

The mean ␬ for the intraobserver reliability of the classification of type 1 and type 2 fractures was 0.846 (range, 0.795-0.933; P ⬍ .05), which can be classified as substantial to almost perfect agreement according to Landis and Koch.15 The multirater ␬ for interobserver reliability was 0.449 (P ⬍ .001), which represents moderate but significant agreement according to Landis and Koch. DISCUSSION The results of our study show that displaced apparently isolated radial head fractures are relatively uncommon, with as few as 6% and no more than 15% of radiographically visible, apparently isolated partial radial head fractures in the practice analyzed, according to the criteria of Broberg and Morrey6 (Figure 1). It is not possible to be more precise than this, because we did not use computed tomography or routine operative inspection to evaluate displacement more precisely. The lack of such a gold standard for diagnostic accuracy is partly because of the retrospective nature of the study but, more importantly, because the vast majority of these fractures are treated nonoperatively, making extra cost, radiation, or operative risk unnecessary and unadvisable. To study the accuracy of radiography further for assessing displacement of stable, apparently isolated partial articular fractures of the radial head, we plan to study fractures with displacement sufficient to consider operative treatment (borderline fractures between Mason type 1 and type 2 or Mason type 2 fractures in this study) using 3-dimensional computed tomography. A gold standard is not necessary to establish that expert observers disagree on fracture displacement to a moderate degree, as we have done. This relatively large series of stable, apparently isolated partial articular fractures of the radial head establishes that the prevalence of displacement, as defined on radiographs via the criteria of Broberg and Morrey, is 15% or less. Considering that these data come from a practice to which patients with displaced fractures are often referred specifically for consideration for surgery, the incidence of displacement of radiographically visible, apparently isolated partial fractures of the radial head may be even smaller in general. We chose to focus on isolated, or apparently isolated, partial articular fractures of the radial head, because they are inherently stable, with intact periosteum, and straightforward to repair with no metaphyseal bone loss, impaction, or missing fragments. These fractures are not important to elbow and forearm stability. Contrast this with the widely displaced,

Doornberg et al

605

unstable partial articular fracture associated with a terrible-triad fracture-dislocation (dislocation with fractures of the radial head and coronoid process). Such a fracture is comminuted and unstable, with little or no intact periosteum, metaphyseal bone loss, and missing or deformed articular fragments, and is difficult to repair, but very important to the stability of the elbow. The contrast between stable, isolated slightly displaced fractures and unstable markedly displaced fractures that are part of a complex injury pattern is substantial and critical. This distinction is probably more important than that between whole-head and partial-head fractures, between comminuted and noncomminuted fractures, and so on, and it may be the most important determinant of the outcome of radial fractures. Additional research is needed to clarify and confirm these concepts. Stable, isolated fractures of the radial head are usually nondisplaced or minimally displaced (a minimum of 85% of radiographically visible fractures and 89% of all apparently isolated radial head fractures in this series). Displaced fractures are usually associated with other fractures or ligament injuries of the elbow or forearm (90/101 fractures in this series, considering the excluded fractures [89%]). Some authors have cautioned that complex fractures of the entire radial head (Mason type 3) are always or nearly always part of a more complex injury pattern7 and that the surgeon should be sure not to overlook associated injuries. A similar caution is probably warranted in the treatment of partial-head fractures that are displaced and unstable. Only 7 of 49 fractures classified as definitely displaced partial-head fractures (14%) in this series were considered isolated fractures. We considered these fractures as isolated based on clinical and radiographic criteria, but it is possible that some of these fractures had associated undiagnosed ligament injuries. For this reason, we refer to these fractures as apparently isolated throughout this article. Itamura et al11 found associated ligament and bone injuries in association with most apparently isolated displaced fractures. Care should be taken not to overlook an associated injury to the forearm or elbow in the presence of any displaced fracture of the radial head, whether it involves part of the head or the entire head. Our data suggest that individual observers can consistently diagnose displacement of isolated partial radial head fractures on radiographs, but they often disagree with each other regarding the presence or absence of displacement. Given that the design of the study would tend to exaggerate disagreement and inconsistency, the finding of moderate interobserver agreement is actually encouraging. On the other hand, the relatively low incidence of displacement must be considered when interpreting radiographs in the clinical setting.

606

Doornberg et al

J Shoulder Elbow Surg September/October 2007

Figure 1 Anteroposterior radiographs of stable, isolated partial articular fractures of radial head. A, Nondisplaced fracture (Mason type 1 according to Broberg and Morrey modification). B, Borderline displaced fracture (borderline between Mason type 1 and type 2 according to Broberg and Morrey modification). C, Displaced fracture (Mason type 2 according to Broberg and Morrey modification). D, Most displaced isolated partial articular fracture in series. The possibility of an undetected ligament injury or fracture is likely very high for this type of fracture.

The relative infrequency of a diagnosis magnifies the weaknesses inherent in any diagnostic test. This is well expressed in the concepts of positive and nega-

tive predictive values and likelihood ratios, where the prevalence of the condition or pretest odds of the diagnosis being present is factored into the formula

Doornberg et al

J Shoulder Elbow Surg Volume 16, Number 5

and indicates a substantially lower probability of the diagnosis being present, even if the test is interpreted as being positive. In other words, if the disease is unlikely in the population being tested, even a relatively limited rate of false positives will become more important, because for any given diagnostic test, a false positive will be encountered more commonly than a true positive. Our study could not evaluate the diagnostic characteristics of radiographs for diagnosing displacement of partial radial head fractures, because we lacked a gold standard, such as observation of all of the fractures at operation; however, we did find greater interobserver variability than intraobserver variability in the radiographic diagnosis of fracture displacement. This suggests the potential for moderate discrepancy between true displacement of greater than 2 mm and the diagnosis of such on radiographs. The drawbacks of such a discrepancy may be magnified by the relative infrequency of substantial displacement of isolated partial radial head fractures, with the result being that displacement may be overdiagnosed on radiographs—that is, there would be a higher rate of false-positive diagnosis of displacement. The best treatment for isolated displaced fractures of the radial head is disputed. Satisfactory results have been reported in the literature with both open reduction–internal fixation8,12,13,18,22 and nonoperative treatment.23 Recent data from Sweden, documenting good or excellent long-term results with nonoperative treatment of displaced partial fractures of the radial head in over 75% of patients by use of a very strict rating scale, bring the role of operative treatment into question.10 A report from the same center focused specifically on the stable, isolated, displaced partial articular fracture of the radial head and noted secondary resection in 6 of 49 patients (12%) for unclear reasons but nearly normal final flexion (137°) and flexion contracture (3°), with 40 patients (82%) being completely without symptoms.1 In the absence of well-executed prospective, randomized trials demonstrating a benefit for operative treatment of displaced partial radial head fractures, the relative infrequency of displacement, the unreliable criteria for radiographic diagnosis of displacement, and the good results of nonoperative treatment of displaced fractures all favor nonoperative treatment for Broberg and Morrey modified Mason type 1 and 2 fractures. We prefer nonoperative treatment, unless the fracture fragments can be shown to restrict forearm rotation on physical examination, a situation that we believe is very uncommon. A clinical trial comparing operative and nonoperative treatment is warranted but may be difficult to execute, given the relative infrequency of isolated, displaced partial radial head fractures and the frequent reluctance of

607

patients to leave the decision to operate to chance (random assignment). More sophisticated imaging techniques, such as computed tomography, may help patients and surgeons decide on the best mode of treatment, but additional research is necessary. REFERENCES

1. Akesson T, Herbertsson P, Josefsson PO, Hasserius R, Besjakov J, Karlsson MK. Primary nonoperative treatment of moderately displaced two-part fractures of the radial head. J Bone Joint Surg Am 2006;88:1909-14. 2. Andersen DJ, Blair WF, Steyers CM Jr, Adams BD, el-Khouri GY, Brandser EA. Classification of distal radius fractures: an analysis of interobserver reliability and intraobserver reproducibility. J Hand Surg [Am] 1996;21:574-82. 3. Audige L, Bhandari M, Hanson B, Kellam J. A concept for the validation of fracture classifications. J Orthop Trauma 2005;19: 401-6. 4. Bhandari M, Montori VM, Swiontkowski MF, Guyatt GH. User’s guide to the surgical literature: how to use an article about a diagnostic test. J Bone Joint Surg Am 2003;85:1133-40. 5. Broberg MA, Morrey BF. Results of delayed excision of the radial head after fracture. J Bone Joint Surg Am 1986;68:669-74. 6. Broberg MA, Morrey BF. Results of treatment of fracturedislocations of the elbow. Clin Orthop Relat Res 1987: 109-19. 7. Davidson PA, Moseley JB Jr, Tullos HS. Radial head fracture. A potentially complex injury. Clin Orthop Relat Res 1993:224-30. 8. Esser RD, Davis S, Taavao T. Fractures of the radial head treated by internal fixation: late results in 26 cases. J Orthop Trauma 1995;9:318-23. 9. Ewald FC, Simmons ED, Sullivan JA, Thomas WH, Scott RD, Poss R, et al. Capitellocondylar total elbow replacement in rheumatoid arthritis: long-term results. J Bone Joint Surg Am 1993;75:498507. 10. Herbertsson P, Josefsson PO, Hasserius R, Karlsson C, Besjakov J, Karlsson M. Uncomplicated Mason type-II and III fractures of the radial head and neck in adults. A long-term follow-up study. J Bone Joint Surg Am 2004;86:569-74. 11. Itamura J, Roidis N, Mirzayan R, Vaishnav S, Learch T, Shean C. Radial head fractures: MRI evaluation of associated injuries. J Shoulder Elbow Surg 2005;14:421-4. 12. Khalfayan EE, Culp RW, Alexander AH. Mason type II radial head fractures: operative versus nonoperative treatment. J Orthop Trauma 1992;6:283-9. 13. King GJ, Evans DC, Kellam JF. Open reduction and internal fixation of radial head fractures. J Orthop Trauma 1991;5:21-8. 14. Kocher MS, Zurakowski D. Clinical epidemiology and biostatistics: a primer for orthopaedic surgeons. J Bone Joint Surg Am 2004;86:607-20. 15. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-74. 16. Mason ML. Some observations on fractures of the head of the radius with a review of one hundred cases. Br J Surg 1959;42: 123-32. 17. Morgan SJ, Groshen SL, Itamura JM, Shankwiler J, Brien WW, Kuschner SH. Reliability evaluation of classifying radial head fractures by the system of Mason. Bull Hosp Jt Dis 1997;56: 95-8. 18. Pearce MS, Gallannaugh SC. Mason type II radial head fractures fixed with Herbert bone screws. J R Soc Med 1996;89: 340P-4P. 19. Sidor ML, Zuckerman JD, Lyon T, Koval K, Cuomo F, Schoenberg N. The Neer classification system for proximal humeral fractures. An assessment of interobserver reliability and intraobserver reproducibility. J Bone Joint Surg Am 1993;75:1745-50.

608

Doornberg et al

20. Siebenrock KA, Gerber C. The reproducibility of classification of fractures of the proximal end of the humerus. J Bone Joint Surg Am 1993;75:1751-5. 21. Siegel S, Castellan JN. Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill; 1988.

J Shoulder Elbow Surg September/October 2007

22. Van Glabbeek F, Van Riet R, Verstreken J. Current concepts in the treatment of radial head fractures in the adult. A clinical and biomechanical approach. Acta Orthop Belg 2001;67:430-41. 23. Weseley MS, Barenfeld PA, Eisenstein AL. Closed treatment of isolated radial head fractures. J Trauma 1983;23:36-9.