Treatment of displaced fractures of the ankle has evolved from routine closed reduction with immobilization in a cast to open reduction and internal fixation for persistently displaced fractures. Lauge-Hansen described the most common patterns of ankle fractures and the mechanisms of injury for each pattern. He emphasized the closed treatment of fractures of the ankle. Subsequent long-term follow-up studies have demonstrated that patients with persistent displacement of the fracture had poorer long-term results than those without persistent displacement and patients with pronation-lateral rotation injuries had poorer results than those with supination-lateral rotation injuries9. Yablon et al. also noted that residual displacement of the fracture led to a poor result, and they found the lateral malleolus to be the key to reduction of the ankle joint. On the basis of the findings of these and other studies, fractures of the ankle, especially pronation-lateral rotation injuries, with residual displacement are treated operatively by most surgeons3,4,9,13,16,17.
In a classic biomechanical study, Ramsey and Hamilton demonstrated a dramatic decrease in the contact area between the tibia and talus with relatively small displacement. Subsequent biomechanical studies, however, have revealed mixed results. For example, the findings in one study15 concurred with those in the study by Ramsey and Hamilton, whereas other studies have demonstrated little or no difference in contact pressures with lateral displacement of the talus5,10,22. Those investigations were limited to the effect of lateral shift of the talus on contact pressures or contact areas in the ankle. When Curtis et al. studied a combination of fibular shortening and external rotation, they found a decrease in the tibiotalar contact area. All of these previous biomechanical studies involved certain non-physiological testing conditions, and none addressed the effect on contact pressure of a combination of malrotation or shortening of the fibula and lateral displacement.
Currently, a lateral shift of the talus of two millimeters or more is a major indication for the operative reduction of a fracture of the ankle4,9,13,16,17. Some authors believe that any visible lateral displacement of the lateral malleolus is unacceptable13. This approach, relying on a readily identifiable radiographic parameter, involves the assessment of only one of the three possible modes of displacement of the lateral malleolus. Of the few studies that have addressed the effect of fibular shortening, one1 led the authors to conclude that no shortening of the fibula is acceptable and two others7,17, that two millimeters of shortening is acceptable. To our knowledge, no one has assessed the effect of malrotation on the end result.
The purpose of the present study was to evaluate the effect of fibular shortening, malrotation, and lateral shift on contact pressures in the tibiotalar joint in a pronation-lateral rotation fracture malunion model. The three deformities were assessed individually and in combination to reflect their effect on contact pressures in the joint more accurately.
*Although none of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article, benefits have been or will be received, but are directed solely to a research fund, foundation, educational institution, or other non-profit organization with which one or more of the authors is associated. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was AO North America.
†Department of Orthopaedic Surgery, University of Southern California, 1200 North State Street, GNH 3900, Los Angeles, California 90033.
‡4233 Sepulveda Boulevard, Sherman Oaks, California 91403.
§2400 South Flower Street, Los Angeles, California 90007.
#2921 South La Cienega Avenue, Suite A, Culver City, California 90232.
Nine fresh-frozen specimens were disarticulated through the knee joint. The specimens were obtained from six women and three men, who had been an average of sixty-eight years old (range, sixty to seventy-five years old) at the time of death. Radiographs were made to rule out preexisting abnormalities of the ankle or foot. All subcutaneous tissues, except for the proximal and distal tibiofibular ligaments and the interosseous membrane, were stripped from the knee to the level of the ankle joint. Approximately five millimeters of the tibial plateau was resected transversely with a saw. An axial loading device with a five-centimeter-long intramedullary rod was cemented onto the surface of the tibial plateau, and fixation of the device was supplemented with two cancellous-bone screws. The loading device served as a link to the materials testing machine (model 8521; Instron, Canton, Massachusetts). In contrast to previous studies5,7,10,14,15,18,22 in which load was applied to the tibia through the mid-portion of the diaphysis, load was applied proximal to the fibula, through the tibial plateau.
The specimen was mounted in the materials testing machine with the plantar aspect of the foot in contact with the foot-plate, which rested on a bearing to prevent the development of rotational or shear forces. Load was applied with the ankle in the neutral position, as it was in all of the previous studies5,10,15,22 of contact pressures in the ankle. One of those studies22 demonstrated significant changes (p < 0.05) in the contact pressure in the ankle joint when the position of the subtalar joint changed. The subtalar joint was placed in approximately 5 degrees of valgus, and two large cancellous-bone screws were inserted from the posterior aspect of the calcaneus into the body of the talus to prevent the subtalar joint from migrating into eversion when load was applied. Five degrees of valgus was selected because it is the position of the subtalar joint during the stance phase of gait when the ankle joint is subjected to maximum load20.
Contact area and contact pressure were quantitated with low-grade pressure-sensitive film (Prescale film; Fuji Photo Film, Tokyo, Japan), which measures pressures from two to seven megapascals. The film was calibrated by applying known contact pressures to a sample of film that had been placed between a flat-ended steel rod and a flat plate, the surfaces of which had been ground to within 2.5 micrometers. The film consists of a pressure-sensitive paper and a registering paper. All measurements were made with the pressure-sensitive film on top of the registering film. The film was scanned, with use of a Howtek scanner (Scan Master 3 Plus; Hudson, New Hampshire) within five hours after exposure, as recommended by the manufacturer, and Sigma Scan image measurement software (version 1.2; Jandel Scientific, San Rafael, California) was used to process images at 300 dots per inch with a sixty-four-level gray scale. Little artifact was noted, but any transverse, linear area corresponding to a crease in the film was subtracted from the images. The gray-scale densities were converted to measurements of pressure (in megapascals) with use of a fourth-order polynomial equation that was determined by non-linear curve-fitting of the calibration curve. After exposure of the film, the images were scanned and the data were divided according to nine quadrants over the dome of the talus.
The film was cut to the shape of the talar dome and was sealed between two 0.05-millimeter-thick polyethylene sheets (3M Tape; 3M, Minneapolis, Minnesota) to prevent damage caused by joint fluids. An anterior capsulotomy was performed, the ankle joint was distracted slightly, and the film packet was inserted over the dome of the talus, avoiding any movement of the tibiotalar articulation to prevent shear artifact. The entire anterior aspect of the joint was marked on the film to serve as a reproducible reference for the dome of the talus. Only the pressure over the dome of the talus was measured, as that is where most axial loading occurs, and contouring the film over the medial and lateral aspects of the talus would have led to unacceptable artifact. For each testing condition, load was applied at a rate of twenty-five newtons per second to a load of 700 newtons; this load was maintained for thirty seconds, and then the specimen was unloaded at a rate of twenty-five newtons per second. The duration of loading was similar to that recommended by McKellop et al., who demonstrated that the color density on pressure-sensitive film was affected by the rate and duration of the applied load. A load of 700 newtons was selected because it represents a body weight of seventy kilograms. Although the approximate maximum load in the ankle during walking on a level surface can approach four times the body weight21, we selected a lower load to minimize the risk of plastic deformation of bone or soft tissues. The ankle is subjected to a maximum load during approximately 50 to 70 per cent of the stance phase of gait, at which time the ankle is in neutral or mild dorsiflexion21; this position corresponds to the neutral position of the ankle in the present study.
A fracture-displacement device was developed that could displace the lateral malleolus with three degrees of freedom: rotation, shortening, and lateral shift. The device was secured to the fibula with use of four-hole AO small-fragment dynamic compression plates (Synthes, Paoli, Pennsylvania) that were welded to the proximal and distal ends of the device. The distal four-hole plate was contoured to fit the distal aspect of the fibula and so that the fibular osteotomy would be created approximately forty-five millimeters proximal to the level of the ankle joint. The device was secured to the fibula with four 3.5-millimeter cortical-bone screws proximally and distally; before each screw was inserted, polyurethane was injected into the hole to augment fixation. The device was anchored to the distal aspect of the fibula at a point where the contour of the distal plate had the best fit against the cortex. The distance between the site of the osteotomy and the ankle joint was measured in each specimen.
The fracture-displacement device allowed for lateral displacement of zero to six millimeters in two-millimeter increments, shortening of zero to six millimeters in two-millimeter increments, and external rotation of 0 to 15 degrees in 5-degree increments (Fig. 1). Baseline values for joint contact area and contact pressure were determined with the device secured to the fibula. The osteotomy then was performed, with resection of one centimeter of bone to allow for displacement of the fibula. Baseline values for contact area and contact pressure were determined again to ascertain whether they had changed as a result of the osteotomy and resection of bone. The soft tissues in each specimen then were sectioned to replicate a pronation-lateral rotation fracture. The deep fibers of the deltoid ligament, the anterior-inferior tibiofibular ligament, and the interosseous membrane were divided, to the level of the osteotomy site, and the posterior aspect of the joint capsule also was divided. We disrupted only the deep fibers of the deltoid ligament because division of the entire ligament (deep and superficial fibers) in two pilot specimens led to severe instability and dislocation of the joint. The contact area and contact pressure were measured again after ligamentous disruption.
The testing protocol involved the measurement of contact area and contact pressure after increases in each of the three displacement variables (two, four, and six millimeters of lateral fibular shift; 5, 10, and 15 degrees of external rotation; and two, four, and six millimeters of shortening). Each of these nine conditions was tested individually. In addition, the variables were paired according to small, medium, or large displacement (two millimeters of lateral shift and 5 degrees of external rotation, four millimeters of lateral shift and 10 degrees of external rotation, and six millimeters of lateral shift and 15 degrees of external rotation, with similar pairs for shortening and external rotation and similar pairs for lateral shift and shortening). Also, three tests were done with a combination of the three displacement variables (two millimeters of both lateral shift and shortening with 5 degrees of external rotation, four millimeters of both lateral shift and shortening with 10 degrees of external rotation, and six millimeters of both lateral shift and shortening with 15 degrees of external rotation). After completion of each series of tests with three progressively greater displacements, the specimen was returned to its non-displaced setting and the baseline contact area and contact pressure were determined again. The protocol resulted in thirty-one tests on each specimen (twenty-one tests with displacement and ten tests to determine baseline values). The sequence of the testing conditions was random in order to minimize its effect on the results.
The mean contact pressure in each of the nine quadrants of the talar dome was calculated under each of the thirty-one testing conditions. The mean baseline values for the quadrants were compared with use of a one-way analysis of variance. A paired t test was used to compare the mean contact pressure for each quadrant under each of the twenty-one displacement conditions with the corresponding baseline measurement. Each p value is reported individually. The total contact area for each of the thirty-one testing conditions was also calculated for each specimen. The baseline total contact areas were compared with use of one-way analysis of variance. The total contact area for each of the displacement conditions was compared, with use of a paired t test, with the baseline total contact area.
The statistical power of the study, to detect a difference of one megapascal or more, was on the order of 90 per cent. For example, in order to detect a difference of at least one megapascal, with a standard deviation of 0.8 and a significance level of 0.05, nine specimens were needed to achieve a power of 90 per cent.
There was a more uniform distribution of contact pressure (range, 1.3 to 2.5 megapascals) over the nine quadrants of the talar dome in the intact specimens than under any of the displacement conditions (Figs. 2-A, 2-B, and 2-C). No change was noted between the baseline contact pressures in the intact specimen and those measured after only the osteotomy had been performed or after ligamentous disruption. No significant changes could be detected, in any quadrant, between the baseline contact pressures in the intact specimens and the baseline contact pressures that were measured after each series of three displacement conditions.
On displacement testing, the highest contact pressures were found in the mid-lateral quadrant of the talar dome (Fig. 2-A). All of the mean pressures were greater than the baseline values in this quadrant. The highest contact pressures were found in association with maximum shortening of the fibula (4.1 megapascals), followed by the combination of maximum shortening and lateral shift (3.9 megapascals), the combination of maximum shortening and external rotation (3.8 megapascals), and the combination of maximum shortening, lateral shift, and external rotation (3.4 megapascals). Almost every displacement testing condition resulted in a significant increase in the contact pressure in the posterolateral quadrant (Fig. 2-A). In the mid-posterior, central, and mid-anterior quadrants, the contact pressures remained relatively unchanged from the baseline values (Fig. 2-B). All of the displacement testing conditions resulted in decreased contact pressures in the medial quadrants, and most of these decreases were significant. In general, increases of each type of displacement (lateral shift, external rotation, or shortening of the fibula) yielded corresponding increases in the contact pressures in the posterolateral and mid-lateral quadrants of the joint, where peak contact pressures were recorded (Fig. 2-A).
Despite the changes in contact pressures, we found no significant difference between any of the baseline values for total contact area and the total contact area measured under any of the displacement conditions.
The site of the osteotomy was an average of 53.4 ± 7.9 millimeters from the ankle joint.
In their classic biomechanical study, Ramsey and Hamilton assessed the contact area of the ankle joint with use of a carbon black transference technique. They found an average 42 per cent reduction in the tibiotalar contact area with one millimeter of lateral displacement of the talus. However, there were non-physiological aspects of their model. For example, the distal part of the fibula was completely resected, and the talus was displaced with use of metallic spacers between the talus and the medial malleolus. In a subsequent study, Moody et al. evaluated the effect of fibular and talar displacement on contact areas and contact pressures in the ankle joint. Using pressure-sensitive film, they found a 50 per cent reduction in the contact area, along with a linear increase in average peak pressures, with only one millimeter of talar displacement. They also displaced the talus by placing premeasured spacers between the medial malleolus and the talus. Curtis et al. reported changes in the contact area with two millimeters of shortening, alone and in combination with 30 degrees of external rotation of the fibula. They did not measure contact pressures or evaluate the effect of lateral shift of the talus or fibula. Other authors, however, have not been able to reproduce these results. Using a supination-lateral rotation model, Clarke et al. found no difference in the contact areas in the joint with lateral shift of the distal end of the fibula. Vrahas et al. used a pronation-lateral rotation fracture malunion model and found no difference in the contact pressures in the joint with one to four millimeters of lateral talar shift. Kimizuka et al. noted similar findings when they evaluated contact areas. The lack of agreement among previous studies may be due to the fact that only contact area was assessed in some studies, that the talus was displaced with washers in the medial joint in some studies, and that only lateral shift of the talus or fibula was assessed in some studies5,7,10,15,18,22. We believe that we are the first to evaluate the individual and combined effects of fibular shortening, lateral displacement, and external rotation. In addition, it is unclear whether a physiological load was applied to the fibula in previous studies, as the load was applied through the mid-portion of the tibial diaphysis. In the present study, load was applied through the tibial plateau in an attempt to ensure a more physiological load on the fibula.
We found significant increases in the contact pressures in the mid-lateral and posterolateral quadrants of the talar dome in association with most of the displacement conditions. The highest peak contact pressures were found in the mid-lateral quadrant in association with maximum shortening of the fibula, followed by the combination of maximum shortening and lateral shift, maximum shortening and external rotation, and maximum shortening, lateral shift, and external rotation. Increases in each of the three displacement parameters yielded corresponding (although not consistently significant) increases in peak contact pressures in the mid-lateral and posterolateral quadrants, where the peak contact pressures were found. The finding of increased pressures in the lateral quadrants of the talar dome corresponds to our clinical observation that degenerative changes develop first in the lateral aspect of the joint in patients who have a malunited pronation-lateral rotation fracture of the ankle. Therefore, we concluded that, in addition to lateral shift, fibular shortening and malrotation should be assessed when determining whether operative reduction is needed for a pronation-lateral rotation fracture of the ankle and when assessing the reduction intraoperatively. This finding may be particularly important when there is substantial comminution, as the length and rotation of the fibula may be more difficult to determine. Furthermore, the significant changes in contact pressures that we found with as little as two millimeters of shortening or lateral shift of the fibula or 5 degrees of external rotation suggest that pronation-lateral rotation injuries should be reduced to within these values.
For each testing condition, we applied a load of 700 newtons to the specimen in order to approximate a body weight of seventy kilograms. Although previous studies have documented that peak loads through the ankle joint can approach four times body weight during walking on a level surface21, we chose to approximate body weight because we were concerned about damaging the fixation of the fracture-displacement device to the fibula. To minimize the possibility of damage, we augmented the fixation of each screw to the fibula with a small amount of polyurethane. We also altered the order of tests on each specimen to eliminate the effect of plastic deformation of bone or ligamentous structures as a result of sequential testing conditions; we found no significant difference in the contact pressures among the specimens. In addition, there were no differences among the baseline values that were measured after each series of displacement conditions; this suggests that little or no plastic deformation of the bone or soft tissue occurred during the course of testing.
The fracture-displacement device was contoured such that the level of the fracture would be approximately forty-five millimeters proximal to the ankle joint. In a previous biomechanical study, Boden et al. showed that a syndesmotic screw was not necessary to secure a fracture that had occurred within 4.5 centimeters proximal to the level of the ankle joint, despite disruption of the deltoid ligament, when the fibula had been rigidly fixed with a plate. In the present study, the site of the fracture ranged from forty-six to sixty-seven millimeters (average, fifty-three millimeters) proximal to the ankle joint. Despite the fact that the fractures were created proximal to the so-called safe level that was described by Boden et al., we do not believe a syndesmotic screw was necessary for stabilization as we found no significant difference between the contact pressures in the intact specimen and those measured after creation of the osteotomy and ligamentous disruption.
In two pilot specimens, we disrupted the entire deltoid ligament (both superficial and deep fibers) to simulate the pronation-lateral rotation injury, but the joints became so unstable that they dislocated when load was applied. Therefore, in the current study, we disrupted only the deep portion of the deltoid ligament, as other investigators have noted that this portion plays the most important role in the prevention of lateral shift of the talus. Sarrafian stated that the deep portion of the deltoid is the strongest component of the ligament and that it prevents lateral shift of the talus. Close demonstrated the importance of the deep fibers of the deltoid ligament in limiting lateral talar shift in a cadaveric model. Harper evaluated the relative importance of the superficial and deep portions of the deltoid ligament in a cadaveric model and concluded that the deep portion is a secondary restraint against lateral talar shift while the lateral malleolus and lateral ligaments are the primary restraints. Michelsen et al. evaluated the effect of sectioning the superficial and deep fibers of the deltoid ligament in a supination-external rotation ankle fracture model. They found the deltoid ligament to be the major stabilizer of the ankle and concluded that the deep portion provided most of the stability. Therefore, we believe that we used an acceptable model, as the deep fibers of the deltoid ligament, the major stabilizer, were divided and the superficial fibers were preserved in order to prevent dislocation of the ankle.
In conclusion, we evaluated the individual and combined effects of fibular shortening, lateral shift, and external rotation on contact pressures in the ankle joint in a pronation-lateral rotation ankle fracture malunion model. We found that all three variables (alone and in combination), especially shortening, increased contact pressures. The findings of the present study support the belief that these fractures should be reduced anatomically, as any fibular displacement that we measured, whether it was shortening, external rotation, or lateral displacement, increased contact pressures in the ankle joint.