New classification of lunate fossa fractures of the distal radius

Background A die-punch fracture is a depression fracture of the lunate fossa of the distal radius. We propose a morphological classification of die-punch fractures that includes five types: center depression fractures, vertical depression fractures, volar depression fractures, dorsal depression fractures, and double die-punch fractures. Methods The radiographs of 112 die-punch fractures treated between January 2005 and January 2015 were retrospectively reviewed. The clinical images were examined independently for two rounds by six orthopedists with different clinical experiences: two residents, two attending physicians, and two consultants. A category-specific kappa score and a kappa score for more than two observers were analyzed. We used Cohen’s kappa to test intraobserver variation. Results The kappa score for interobserver reliability was 0.69 for the first round and 0.70 for the second round. The intraclass correlations were 0.65 and 0.63, respectively. Intraobserver reproducibility using Cohen’s kappa test was satisfactory. All of the results indicated a kappa value >0.4, suggesting good agreement within, as well as between, observers. Conclusions The outcome was assessed using kappa statistics, which showed good interobserver reliability and intraobserver reproducibility.


Background
Distal radius fractures are common upper limb injuries in persons of any age. Many classifications have been proposed for distal radius fractures [1]. The AO [2], Universal [3], and Fernandez [4] classifications are among those most frequently used. However, there are no detailed descriptions of the characteristics of the die-punch fracture in the existing classifications. A die-punch fracture is a depression fracture of the lunate fossa of the distal radius that is caused by a vertical load through the lunate. We believe that the previous definition of a die-punch fracture is not sufficiently comprehensive, making it easy to forget or ignore its diagnosis and treatment. In other cases, the die-punch fracture is in a state that is difficult to reset, such as when it is not limited to the lunate fossa. Therefore, redefinition of die-punch fractures can reduce the lack of attention to rare fractures, improve the accuracy of the reduction, and thus improve the treatment effect.
Five types of lunate fossa depression fracture have been found in our clinical practice, including center depression fractures, vertical depression fractures, volar depression fractures, dorsal depression fractures, and double die-punch fractures. This new classification aims to indicate the severity of the injury and thus project a prognosis based on the presumed complexity of the bone injury. Orthopedic surgeons use classification systems as a guide for treatment and prognosis as well as for comparing results in clinical studies. It is therefore of paramount importance that these systems are reliable and reproducible. All classification systems should be subjected to reliability and reproducibility testing prior to acceptance into widespread use.
The aim of this study was to establish a classification of die-punch fractures of the distal radius based on plain radiographic findings. Both intraobserver reproducibility and interobserver reliability of this classification were examined.

Patients
This study was designed according to previously proposed guidelines for reliability studies of fracture classification systems [5,6]. The clinical images of 112 patients with die-punch fractures treated at our institution from January 2005 to January 2015 were retrospectively reviewed. Patients with the following conditions were excluded: distal radius fractures not involving the lunate fossa, fractures involving other than the distal radius, and patient age <18 years. No radiographs were excluded because of poor quality.

Clinical images
Standard anteroposterior and lateral radiographs and computed tomography (CT) images of the distal radius obtained within 6 days of the injury were reviewed. The anteroposterior radiographs were obtained with the patient seated, arm abducted, elbow flexed, and hand and wrist laid flat on the radiography cassette. The wrist was in neutral position, and the vertical X-ray beam was centered on the wrist. Lateral radiographs were obtained in the true lateral position. A well-positioned lateral view shows the pisiform superior to the dorsal cortex of the distal pole of the scaphoid and the volar cortex of the capitate.

Classification method
The clinical images were examined independently for two rounds by six orthopedists with different clinical experiences: two residents, two attending physicians, and two consultants. The die-punch fractures were classified into five types: center depression fractures ( Fig. 1), vertical depression fractures (Fig. 2), volar depression fractures (Fig. 3), dorsal depression fractures (Fig. 4), and double die-punch fractures (Fig. 5).

Statistical analysis
Statistical analyses were performed using SPSS 19 software (IBM, Armonk, NY, USA) and the associated IRR package [7]. A category-specific kappa score and a kappa score for more than two observers were analyzed [8]. We used Cohen's kappa to test intraobserver variation [9]. A kappa score of 1 indicates perfect agreement, and a score of zero indicates that the variation in agreement can be explained purely by chance. Analysis of variance and χ 2 tests were carried out where appropriate to compare the five fracture types. A value of P < 0.05 was considered to indicate statistical significance.

Results
The mean age of our patients was 59.3 years (range 18-87 years). The patient groups of different fracture classifications did not differ significantly in regard to age, sex, weight, height, body mass index, osteoporosis, or smoking status ( Table 1) The kappa score for interobserver reliability was 0.69 for the first round and 0.70 for the second round ( Table 2). The intraclass correlations were 0.65 and 0.63, respectively. Intraobserver reproducibility using Cohen's kappa test is shown in Table 3. All of the results indicated a kappa value of >0.4, suggesting good agreement within, as well as between, observers.

Discussion
The die-punch fracture was first described by Scheck et al. in 1962 as a dorsal fracture fragment of the lunate fossa at the distal radius [10]. The current concept of die-punch fracture is a depression fracture of the lunate fossa of the distal radius. There is no consensus on the classification or treatment guidelines for die-punch fractures. In our practice, five types of die-punch fractures have been found, including center depression fractures, vertical depression fractures, volar depression fractures, dorsal depression fractures, and double diepunch fractures. Among our patients, the dorsal depression fracture was the most frequent classification (44/112, 39 %), which might be related to the volar inclination of the distal radius articular surface. In addition, the dorsal side of the articular surface is higher   than the volar side, which causes the dorsal side of the articular surface to be more prone to injury upon impact. In addition, we found that there was no statistical difference between fracture type and the patient's sex, age, weight, height, body mass index, presence of osteoporosis, or smoking status.
In the AO classification, fractures involving the articular surfaces are classified into types B and C. Type B fractures are usually caused by light trauma and involve only the local articular surface. Although the local articular surface can be comminuted, the entire articular surface and the metaphysis are not comminuted in type B fractures. Accordingly, our study excluded patients whose fractures involved areas other than the distal radius, such as those with metaphysis of the distal radius. Therefore, in our study, the die-punch fracture was restricted to the ulnar half of the articular surface at the distal radius, which is caused by depression of the lunate fossa. In this condition, the radial half of the articular surface at the distal radius remains intact. Thus, the integrity of the distal radius is not completely destroyed. Our classification of the diepunch fracture is based on this definition of lunate fossa depression at the distal radius.
The main advantage of our classification is that observers are able to classify the fracture objectively according to computed tomography faultage parameters. They would not have to offer a judgment based on a subjective evaluation. Based on the principle of the practicality and reproducibility of a fracture categorization, we developed this classification to guide clinical treatment and infer a straightforward prognosis.
The wrist ligament is attached to the edge of the radius joint. Hence, a die-punch fracture would be difficult to reset using closed reduction. Because of the severe instability of volar and dorsal depression fractures, the displacement may reappear with the passage of the time even if the initial displacement is not severe. Moreover, according to Rikli's three-column theory [11], about 40 % of the axial load is transferred by the lunate fossa, so anatomical reduction of the lunate fossa is very important. A non-locking steel plate should be fixed on the side where the replacement is unstable.
For a type I center depression fracture, the reset should be achieved with a small window via a volar or dorsal approach combined with bone grafting. In addition, if the collapse is >5 mm or osteoporosis is present, the joint surface should be buttressed with a steel plate. For type II vertical depression fractures, the use of a volar plate could obtain satisfactory results. For type III volar depression fractures, a modified Henry volar approach can be applied, fixing a steel plate on the palm side. A volar incision could be less affected upon tendon. For type IV dorsal depression fractures, a dorsal approach with fixing a steel plate at the back side is suggested. The multi-angle stability of a locked steel plate makes it possible to achieve fixation of the back side of the bone using a steel plate on the palm side. For type V double die-punch fractures, if fixation of a steel plate on the palm side cannot reset the bone on the back side, the dorsal approach should be applied. To obtain maximum holding force, distal screws placed under the subchondral bone can enhance the stability of the fixation, especially in osteoporotic patients. Type I and V fractures are difficult to treat surgically as they are prone to poor resolution on the joint surface, followed by emergence of traumatic arthritis of the wrist. Hence, it is essential to restore the flatness of the joint surface to the extent possible.