Tuesday, August 25, 2009

Distal Femur Defects Reconstructed With Polymethylmethacrylate and Internal Fixation Devices: A Biomechanical Study


By Anthony D. Uglialoro, MD; Michael Maceroli, BS; Kathleen S. Beebe, MD; Joseph Benevenia, MD; Francis R. Patterson, MD
ORTHOPEDICS 2009; 32:561

Abstract

Benign aggressive distal femur tumors are treated with curettage, adjuvant phenol or argon, and polymethylmethacrylate (PMMA) packing. For large defects, an internal fixation device is added to reduce the fracture risk. The purpose of this study is to compare the strength of locking plates to other fixation devices for stabilization of these defects.

Lateral condyle defects in young, fresh frozen femurs were packed with PMMA and augmented by internal fixation. Three groups of 4 matched pairs of femurs were organized for the following comparisons: (1) stacked Steinmann pins vs crossed screws; (2) stacked pins vs locking plates; and (3) crossed screws vs locking plates. Specimens were subjected to axial load-to-failure testing on an MTS machine.

There was no difference in load-to-failure strength (P=.177) using Steinmann pins or crossed screws. Locking plate constructs were stronger (P=.028) than Steinmann pin constructs. Locking plate constructs were also stronger (P<.001) than crossed-screw constructs. Steinmann pin constructs failed with severe intra-articular fractures; crossed screw constructs failed with bulging of the defects, articular impaction, and minimal fracture propagation. Locking plate constructs failed with extra-articular spiral shaft fractures.


advertisement
Cipka

The distal femur is a common site for many benign bone tumors, such as giant cell tumor of bone. Tumors of this type are typically treated via extended curettage through a large cortical window.1,2 After curettage the local bony architecture is left markedly disturbed and the host bone is vulnerable to postoperative fracture.3,4 Therefore, the lesion is typically packed with polymethylmethacrylate (PMMA) cement. Studies have shown reconstruction with PMMA has improved stability of the repaired bone while still allowing for ease of a second intervention if necessary.5,6 Despite the advantages of curettage and cementation, if a defect occupies >50% of the metaphysis and extends down to the articular surface, internal fixation is typically used to reduce the chances of postoperative fracture.7-9

In the past, various types of devices and configurations have been used to augment the PMMA in post-curettage defects, including stacked Steinmann pins and crossed screws. Studies have compared the biomechanical properties of these reconstructive techniques. In particular, Randall et al10 showed tibial defects reconstructed with Steinmann pins augmenting PMMA displayed significantly greater load to failure and survival when compared to specimens repaired with PMMA alone. Despite these results, other studies have found no significant difference between PMMA and PMMA supplemented with Steinmann pins.11,12 In a pivotal study, Toy et al13 showed that distal femur defects reconstructed with crossed screws augmenting PMMA cement displayed a significant biomechanical advantage over femurs repaired with PMMA alone or PMMA reinforced with Steinmann pins. To date, no study has compared reconstruction with Steinmann pins or crossed screws to locking condylar plates, the current preferred method of internal fixation.

Locking plates are now commonly indicated for internal fixation of peri-articular distal femur fractures, especially for those with metaphyseal communition, providing a rigid, toggle-free, fixed angle construct. Furthermore, locking plates represent an effective, post-curettage internal fixation device to support acrylic cementation in treatment of giant cell tumor. Studies in cadaveric models have demonstrated the structural superiority of locking plates over an unlocked, limited contact–dynamic compression plate (LC-DCP).14 The use of locking plates in distal femur fracture has shown a reduction in fixation failure, as well as less varus collapse.15 In addition, when used for fracture fixation, the locking plate can lie above the bone surface thereby preserving vascular supply of bone and preventing postoperative fractures, a common side effect of friction in LC-DCP reconstruction.14

Despite clinical support for the use of locking plates in bone reconstruction, to date no current biomechanical study has compared the stability of locking plates to Steinmann pins and crossed screws for post-curettage augmentation of PMMA. Our aim in this study is to determine which combination of polymethylmethacrylate (PMMA) and internal fixation device will yield the most biomechanically stable construct.

Materials and Methods

Twelve pairs of matched femora from human cadavera obtained from the Musculoskeletal Transplant Foundation (Edison, New Jersey) were divided into 3 groups each containing 4 pairs of fresh-frozen femora; a total of 8 specimens per testing group. The average age of donors at the time of death was 38 years (range, 21-54 years). The differences in age and gender were accounted for by comparing the post-fixation strength and stiffness of a given specimen to that of its matched counterpart that was internally stabilized with a different experimental fixation device. This allowed the assumption that although the mechanical resistance may be dissimilar if comparing specimens from 2 different donors, the strength and stiffness of a matched pair of femora from a single donor should be nearly identical.

Preparation of Defects

The fresh-frozen femora were thawed at room temperature. The femora were stripped of all soft tissues and radiographs were made to confirm osseous integrity. A high-speed burr (Medtronics, Fort Worth, Texas) was then used to create a defect in the lateral condyle of all specimens, extending from the junction of the metaphyseal and diaphyseal areas down to the subchondral bone. This defect represents the cavity that remains after a giant cell tumor is excised via curettage from the lateral femoral condyle. In the coronal/frontal plane, the defect extended from the edge of the lateral femoral condyle to the trochlear groove (distance x) (Figure 1). In the sagittal plane, the defect extended from the entire anteroposterior width (distance y) of the lateral femoral condyle, leaving a cortical shell intact (Figure 2). Cortical and cancellous bone was removed from the subchondral plate to the epiphyseal/metaphyseal junction (Distance x’). After the defect had been created in each matched pair, the reconstruction method was determined by a random number generator.

Figure 1: The defect extended from the edge of the lateral femoral condyle to the trochlear groove Figure 2: The defect extended from the edge of the lateral femoral condyle to the trochlear groove

Figure 1: In the coronal/frontal plane, the defect extended from the edge of the lateral femoral condyle to the trochlear groove (distance x). The vertical height was equal to this distance (x’). Figure 2: The vertical height was equal to the distance x’. In the sagittal plane, the defect extended from the edge of the lateral femoral condyle to the trochlear groove (leaving a cortical shell intact).

Methods of Fixation

Three groups of 4 matched pairs of femora were organized for the following comparisons: (1) Steinmann pins vs crossed screws; (2) Steinmann pins vs locking plates; and (3) crossed screws vs locking plates. In each respective group, 1 of the matched pairs from a single donor was randomly assigned 1 of 2 different experimental fixation devices, while its matched counterpart received another stabilization method.

For femora reconstructed with Steinmann pins, three 3/16-in threaded intra-medullary Steinmann pins (Zimmer Inc, Warsaw, Indiana) were manually inserted into the medullary canal such that they extended into the diaphysis. The distal ends of the Steinmann pins were fanned out to support the lateral joint surface (Figure 3).

Figure 3A: Post-curettage distal femur reconstructed with Steinmann pins Figure 3B: Gross image of Steinmann pin placement

Figure 3: Lateral radiograph of post-curettage distal femur reconstructed with Steinmann pins without PMMA (A). Gross image of Steinmann pin placement in the lateral femoral condyle (B).

For femora reconstructed with crossed screws (Synthes, Paoli, Pennsylvania), a 3.2-mm bit was used to drill 4 guide holes extending to the opposite cortex. Two holes were directed from the distal aspect of the femur proximally (superiorly) to the contralateral cortex, while the other 2 holes were directed inferiorly toward the opposite medial condyle. Two 4.5-mm cortical screws were placed into the superiorly directed holes, and two 6.5-mm cancellous screws were placed in the inferiorly directed holes. The screws were advanced into the bone until the screw heads were aligned with the missing border of the lateral condyle (Figure 4).

Figure 4A: The distal femur repaired with crossed screws Figure 4B: The distal femur repaired with crossed screws

Figure 4: AP (A) and lateral (B) radiographs of the distal femur repaired with crossed screws.

For femora reconstructed with locking condylar plates (Synthes) the standard, accepted method was used (Figure 5).

After all femora were reconstructed with either Steinmann pins, crossed screws, or locking plates, PMMA was prepared and mixed. The cement was molded into the defect, around the pins, screws, or locking plate and shaped to form the physiologic contour of the lateral femoral condyle. The cement was then given ample time to cure.

Figure 5A: The defect can be seen in the lateral condyle of this femur sample Figure 5B: The distal femur repaired with crossed screws Figure 6: The loading nose was centered to apply an even load to both femoral condyles

Figure 5: AP radiograph (A) and gross image (B) of reconstruction with locking plate. Note that the defect can be seen in the lateral condyle of this femur sample. Figure 6: Distal femur sample reconstructed with a locking plate shown secured in the INSTRON machine. The loading nose was centered to apply an even load to both femoral condyles.

Testing

Each femur was transected 25 cm proximal to the joint line with a hand saw. The transected femora were then placed into a stainless steel fixture and fixed into acrylic cement. All femora were set into the fixture at an angle that placed the transcondylar axis horizontally when mounted in the testing apparatus.

Each femur, set in its steel fixture, was bolted into the load frame of the Instron machine (Instron, Norwood, Massachusetts). The loading nose was centered to apply a physiological load to both femoral condyles (Figure 6). Starting at 0 load and displacement the load was increased by 10 N/sec until the load reached 475 N. The load was then cycled in a sinusoidal pattern between 50 and 900 N for 2000 cycles at 1 Hz. Each femur that survived was loaded at 1 mm/sec under displacement-controlled feedback until failure. For this trial, failure is defined as a sudden drop of 445 N from the maximum observed load (Figure 7, load vs displacement curve). Load and displacement were recorded during testing and the load vs displacement curve was plotted. From these graphs, the load to failure (N) and the stiffness (N/mm) was calculated and recorded for each femur. In addition, the mode of failure was noted by the authors, photographed, and recorded for each specimen.

Figure 7: The force vs displacement curve

Figure 7: The force vs displacement curve illustrates how we determined load to failure. A drop in Force >445 N was considered a failure. The force measurement immediately preceding this drop was determined to be the load to failure.

Statistical Analysis

Within each group, a paired t test was used to compare differences between treatments for load to failure and stiffness.

Results

All three groups of femora survived the initial sinusoidal pattern of 50 to 900 N for 2000 cycles at 1 Hz.

Group 1 consisted of femora reconstructed with PMMA and Steinmann pins versus PMMA and crossed screws. The mean load to failure for femora reconstructed with PMMA and Steinmann pins was 14916±2959 N compared to 11651±3074 N for femora repaired with crossed screws augmenting PMMA (Table 1). There was no significant difference (P=.831) in load to failure between Steinmann pin and crossed screw constructs.

Table 1: Data on Load to Failure and Stiffness

Group 2 compared femora repaired with PMMA and Steinmann pins against those fixed with PMMA and locking plates. Femora reconstructed with Steinmann pins augmenting PMMA failed at an average load of 11728±2724 N compared with 24245±8228 N for the femora reconstructed with locking plates and PMMA (Table 1). For all of the matched pairs, the femur samples reconstructed with locking plates maintained a more stable construct (P=.013) than femora repaired with Steinmann pins.

In group 3, comparing femora fixed with PMMA and crossed screws to those repaired with PMMA and locking plates, the average load to failure for femora reconstructed with crossed screws augmenting PMMA cement was 9880±1130 N while the load to failure for femora repaired with locking plates and PMMA was 22188±3622 N (Table 1). In all pairs of femora, the locking plate reconstructions were significantly stronger (P=.004) than their matched counterparts repaired with crossed screws augmenting cement. Figure 8 displays the load to failure for groups 1-3.

Figure 8: Load to failure for groups 1-3

Figure 8: Load to failure for groups 1-3. Locking plates represent a significantly stronger construct than Steinman pins and crossed screws in groups 2 and 3, respectively.

Within each matched group, the femora reconstructed with locking plates augmenting PMMA (P=.03) had significantly greater stiffness than the contralateral matched pair repaired with Steinmann pins and PMMA (Table 1). However, there was no significant difference in stiffness between locking plates versus crossed screws augmenting PMMA (P=.552). In addition, stiffness values for Group 1 femora, Steinmann pins versus crossed screws, were not significantly different (P=.334).

Table 2: Mode of Failure

Table 2 displays the mode of failure for all femur samples. All 8 femora reconstructed with Steinmann pins and PMMA failed via a severe, intra-articular (intercondylar) fracture (Figure 9). Seven of the 8 femora repaired with crossed screws augmenting PMMA showed an expanded cortex resulting in bulging and impaction of the articular surface with minimal propagation down the shaft (Figure 10). One crossed screw and PMMA construct failed via an extra-articular fracture. Failure of 6 femora reconstructed with locking plates and PMMA resulted in an extra-articular spiral fracture either anterior or posterior to the locking plate screws (Figure 11). Two of the 8 femora in the locking plate group did not show extra-articular fracture. One of these failed via a supracondylar fracture (Figure 12) while the other femur survived mechanical testing without fracture thereby reaching load capacity (~30,000 N) of the mechanical testing machine. In all femora that were fixated with LCP, the articular surface was noted to be intact without any signs of compromise or fracture.

Figure 9: All 8 femora reconstructed with Steinmann pins and PMMA failed Figure 10: Seven of the 8 femora showed bulging

Figure 9: All 8 femora reconstructed with Steinmann pins and PMMA failed via a severe, intra-articular (intercondylar) fracture. Figure 10: Seven of the 8 femora repaired with crossed screws augmenting PMMA showed bulging and impaction of the articular surface with minimal propagation down the shaft.


Figure 11: Extra-articular spiral fracture either anterior or posterior to the locking plate screws Figure 12: One locking plate reconstruction failed via a supracondylar fracture

Figure 11: Failure of 6 femora reconstructed with locking plates and PMMA resulted in an extra-articular spiral fracture either anterior or posterior to the locking plate screws. Figure 12: One of the 8 locking plate reconstructions failed via a supracondylar fracture.

Discussion

Considering the extensive curettage required to treat giant cell tumors, durable reconstruction is needed to prevent postoperative fractures. Polymethylmethacrylate cementation is accepted as the foundation for noncontained giant cell tumor defect reconstruction for both the stability and adjunctive thermal necrotic effect it offers.16 The need for reinforcement in post-curettage defects has been debated in the literature and various methods of post-curettage augmentation have been tested including stacked Steinmann pins and crossed screws. To our knowledge no studies have evaluated the biomechanical advantages offered by locking plate augmentation to PMMA for reconstruction of giant cell tumor defects.

A retrospective study conducted by Bini et al5 found that giant cell tumor patients whose lesions were fixed with threaded Steinmann pins and PMMA did not experience postoperative fractures. Randall et al10 corroborated this conclusion in a biomechanical study of lateral tibial condyles showing that defects reinforced with PMMA and Steinmann pins displayed significantly greater load to failure than those repaired with PMMA alone. More recently however, Murray et al11 found no significant biomechanical differences between femoral condyle defects repaired with PMMA and Steinmann pins or PMMA alone, further escalating the debate about the need for reinforcement in post-curettage defects.

In a biomechanical study using fresh-frozen femur samples, Toy et al13 demonstrated that distal femur defects repaired with PMMA augmented by crossed screws resulted in a stronger reconstruction than PMMA alone or PMMA with Steinmann pins. In this study, large noncontained defects were created in the medial femoral condyles of 20 matched pairs of human femurs. The femora in various groups were tested to find stiffness and load to failure, with failure being defined as a sudden drop of 445 N. In all trials, the PMMA and crossed screws group represented a significant biomechanical advantage over PMMA alone and PMMA reinforced with Steinmann pins. As noted by the authors, the use of matched pairs of femora minimized variability and allowed the focus to remain on comparing the mechanical advantages of each construct. Acknowledging the power of this study, our design was modeled closely on the methods described by Toy et al13 with the addition of PMMA reinforced with femoral locking plates.

Previously acknowledged as an effective mode of fracture fixation, the versatility of locking plates and their application in oncological treatment has been largely anecdotal. A recent study examined 25 patients receiving locking plates for oncological reconstruction. It was shown that 23 of 25 locking plates were intact after a mean follow-up of 18.2 months. These results are encouraging and show clinically, locking plates can provide a reliable and stable option for oncological reconstruction.17 Through biomechanical testing, our aim was to determine if this conclusion is applicable to distal femoral defects resultant from giant cell tumor resection and curettage. Our study was modeled to elucidate the role locking plates could play in the treatment of this benign, locally aggressive bone tumor.13

To our knowledge the present study is the first biomechanical evaluation of distal femoral locking plates used for oncological reconstruction. This study compares PMMA augmented by a distal femoral locking plate with the previously examined crossed-screw and intramedullary Steinmann pin constructs. Fresh-frozen femurs were used to best represent in vivo bone quality, and cross-matched femur pairs allowed us to control for femur quality and size providing a more powerful comparison between the constructs. Defects of the lateral femoral condyle were created to mimic those caused by giant cell tumor excision and curettage. A load-to-failure biomechanical analysis of the lateral femoral condyle was then conducted comparing cross-matched femurs reconstructed with cement and augmented with intramedullary Steinmann pins, crossed-screws or condylar locking plate.

The significant differences in load-to-failure and stiffness between the different constructs validate our hypothesis that locking plate augmented PMMA is biomechanically superior to crossed-screw or Steinmann Pin augmentation. In all pairs of femora, the locking plate reconstructions showed a significantly higher load to failure (P=.004) than their matched counterparts. The locking plates failed at forces 43% and 53% greater than the Steinmann pins and crossed-screws, respectively. Furthermore the fracture pattern differed between the groups. The Steinmann pin reinforced femurs failed through a severe intra-articular fracture. All but 1 of the crossed-screw augmented femurs failed through bulging of the articular surface (expanded cortex) and spiral fracture. One crossed screw construct failed via an extra-articular fracture. Of the femora reconstructed with plates, 6 of 8 failed via an extra-articular spiral fracture anterior or posterior to the locked screws. Of the other 2 locking plate reconstructions, 1 failed via supracondylar fracture while the other did not fail. As noted by Toy et al,13 failure via an extra-articular fracture is more desirable. These fractures often can be repaired with standard open reduction and internal fixation while fractures involving the articular surface are more difficult to salvage often requiring allograft or endoprosthetic reconstruction.

Our explanation as to why the locking plate reconstruction showed a significant advantage over Steinmann pins or crossed screws is because of the additional stiffness the plates provide. As noted by Toy et al,13 the addition of screws crossing the midline in the condylar defect improves stability of the cement mantle thus transferring the load proximally. The multiple locking screws used in the locking plate head essentially act as crossed screws used in Toy’s study. As supported by the extra-articular type fractures seen in the locking plate reinforced femora, this allows the axial compression forces applied to the condyle to be transmitted proximally to the femur shaft that has better bone quality and is undisturbed by tumor and curettage. The superior fixation provided by multiple fixed-angle cortical screws also stabilizes femoral shaft to provide additional coronal stability.

The limitations of this study need to be considered before extrapolating the results for clinical application. For biomechanical testing, the femora were mounted at an angle in the testing apparatus such that the transcondylar axis was oriented horizontally. By orienting the femora as such, the applied force mimicked physiologic load to most specifically test the strength of each construct. Perhaps future biomechanical studies could examine the effects of torsion on these 3 modes of reconstruction to further delve into the response to physiologic pressure. Additionally these results should not be extrapolated to other anatomic sites. Locking plate augmentation of PMMA also has the disadvantage of being a more technically difficult surgery than cement alone or cement augmented with crossed-screws. Furthermore the additional hardware poses greater difficulty if there is recurrence of tumor or the hardware fails and a revision surgery is required. Therefore these data must be considered as biomechanical in nature only. Although the additional strength of the construct and the clinical evidence reported by Virkus et al17 appear to support the routine use of locking plates for reconstruction of giant cell tumor defects, the aforementioned shortcomings should lead practicing surgeons to take a more selective approach in managing giant cell tumor with locking condylar plates.

In this in vitro study, reconstruction of post-curettage distal femora with locking plates and polymethylmethacrylate was biomechanically stronger in axial load testing than Steinmann pins or crossed screws augmenting PMMA. The locking plate constructs were able to withstand compression forces exceeding physiologic levels. This resilient construction may permit more rapid mobilization and prevent postoperative fracture. If the locking plate reconstruction failed, the ensuing fracture is likely to be extra-articular and would preserve the articular surface. Fractures of this type are easier to treat and rehabilitate than the intercondylar fractures associated with other modes of reconstruction. Although not recommended for routine management, locking condylar plates present a significant biomechanical advantage over opposing devices such that LCP should be the preferred reconstruction method when giant cell tumor is complicated by pathological fracture.

References

  1. McDonald DJ, Weber KL. Giant cell tumor of bone. In: Schwartz HS, ed. Orthopaedic Knowledge Update: Musculoskeletal Tumors 2 Rosemont, IL: American Academy of Orthopaedic Surgeons; 2007:133-140.
  2. McGarry SV, Gibbs CP. Giant cell tumor of bone. In Damron TA, ed. Orthopaedic Surgery Essentials: Oncology and Basic Science Philadelphia, PA: Lippincott Williams and Wilkins; 2008:159-164.
  3. Buecker PJ, Gebhardt MC. Are fibula strut allografts a reliable alternative for periarticular reconstruction after curettage for bone tumors? Clin Orthop Relat Res 2007; (461):170-174.
  4. Marcove RC, Weis LD, Vaghaiwalla MR, Pearson R. Cryosurgery in the treatment of giant cell tumors of bone: a report of 52 consecutive cases. Clin Orthop Relat Res 1978; (134):275-289.
  5. Bini SA, Gill K, Johnston JO. Giant cell tumor of bone: curettage and cement reconstruction. Clin Orthop Relat Res 1995; (321):245-250.
  6. Persson BM, Wouters HW. Curettage and acrylic fixation cementation in surgery of giant cell tumors of bone. Clin Orthop Relat Res 1976; (120):125-133.
  7. Leeson MC, Lippitt SB. Thermal aspects of the use of polymethylmethacrylate in large metaphyseal defects in bone. A clinical review and laboratory study. Clin Orthop Relat Res 1993; (295):239-245.
  8. Mjöberg B, Pettersson H, Rosenqvist R, Rydholm A. Bone cement, thermal injury and the radiolucent zone. Acta Orthop Scand 1984; 55(6):597-600.
  9. Weber KL. Surgical treatment of benign bone tumors about the knee. Techniques in Knee Surgery 2003; 2(4):274-285.
  10. Randall RL, Weenig KN, West JR, Johnston JO, Bachus KN. Durability and strength of Steinmann pin augmentation in cemented tibial defects. Clin Orthop Relat Res 2002; (397):306-314.
  11. Murray PJ, Damron TA, Green JK, Morgan HD, Werner FW. Contained femoral defects: biomechanical analysis of pin augmentation in cement. Clin Orthop Relat Res 2004; (420):251-256.
  12. Weiner M, Damron TA, Patterson FR, Werner FW, Mann KA. Biomechanical study of pins in cementing of contained proximal tibia defect. Clin Orthop Relat Res 2004; (419):232-237.
  13. Toy PC, France J, Randall L, Neel MD, Shorr RI, Heck RK. Reconstruction of noncontained distal femoral defects with polymethylmethacrylate and crossed-screw augmentation: a biomechanical study. J Bone Joint Surg Am 2006; 88(1):171-178.
  14. Haidukewych GJ. Innovations in locking plate technology. J Am Acad Orthop Surg 2004; 12(4): 205-212, 2004.
  15. Haidukewych GJ, Ricci W. Locked plating in orthopaedic trauma: A clinical update. J Am Acad Orthop Surg 2008; 16(6):347-355.
  16. Komiya S, Inoue A. Cementation in the treatment of giant cell tumor of bone. Arch Orthop Trauma Surg 1993; 112(2):51-55.
  17. Virkus WW, Miller BJ, Ping CC, Gitelis S. The use of locking plates in orthopedic oncology reconstructions. Orthopedics 2008; 31(5):1.

Authors

Drs Uglialoro, Beebe, Benevenia, and Patterson and Mr Maceroli are from the University of Medicine and Dentistry of New Jersey – New Jersey Medical School, Newark, New Jersey.

Drs Uglialoro, Beebe, Benevenia, and Patterson and Mr Maceroli have no relevant financial relationships to disclose.

The authors thank the Musculoskeletal Transplant Foundation (Edison, New Jersey) and Synthes Inc (Paoli, Pennsylvania) for sponsoring this project. The authors also thank Assimina A. Pelegri, PhD, and her staff at Rutgers University, Department of Mechanical and Aerospace Engineering (Piscataway, New Jersey).

Correspondence should be addressed to: Francis R. Patterson, MD, Department of Orthopedics, Division of Musculoskeletal Oncology, UMDNJ – New Jersey Medical School, 140 Bergen St, ACC Building, Ste D-1610, Newark, NJ 07103.

DOI: 10.3928/01477447-20090624-29

No comments:

Post a Comment