- Email: [email protected]
A biomechanical study on flexible intramedullary nails used to treat pediatric femoral fractures
Journal of Orthopaedic Research 23 (2005) 1315–1320 www.elsevier.com/locate/orthres
A biomechanical study on flexible intramedullary nails used to treat pediatric femoral fractures Jason K. Green *, Frederick W. Werner, Raman Dhawan, Peter J. Evans, Sean Kelley, Dwight A. Webster Department of Orthopedic Surgery, SUNY Upstate Medical University, 3214 Institute for Human Performance, 505 Irving Avenue Syracuse, NY 13210, USA Accepted 21 April 2005
Abstract Flexible intramedullary nails have been indicated to treat femoral fractures in pediatric patients. The purpose of this study was to examine the stability of simulated transverse fractures after retrograde intramedullary flexible nail fixation. Various nail diameter combinations were tested using composite femurs in bending, torsion, and a combined axial/bending test where a vertical compressive force was applied to the femoral head. The cross-sectional percent area fill of the nails within the femurs was also determined. In 4 point bending, the greatest repair stiffness was 12% of the intact stiffness. In torsion, the greatest stiffness was 1% of the intact stiffness for either internal or external rotation. The greatest repair stiffness was 80% of the intact stiffness for a compressive load applied to the femoral head. Nail combinations with single nail diameters greater than 40% of the mid-shaft canal width, as measured from an AP radiograph, prevented the fracture from being reduced and left a posterior gap. Flexible intramedullary nails may be of value in the treatment of pediatric femoral fractures, but care must be taken to insert nails that are correctly sized for the canal and to protect the healing fracture from high torsional and bending loads. 2005 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. Keywords: Pediatric femoral fractures; Flexible intramedullary nail
Introduction Flexible intramedullary nails have been successfully used in treating pediatric femur fractures [4,5,8,11,12]. These have been used in an age group where conservative treatment in a hip spica cast can be difficult on the patient and family due to limited mobility. The reduced immobilization time associated with flexible intramedullary nails is also believed to aid the healing process and eliminate complications of external fixators [4,5]. The optimal diameter and number of nails that should be introduced into the canal to maximize stability is unknown. Although the literature shows that these *
Corresponding author. Tel.: +1 315 464 5228; fax: +1 315 464 6638. E-mail address: [email protected] (J.K. Green).
implants have been successful, a clinical complication was observed by the authors in which a pediatric patient grossly bent a pair of retrograde flexible intramedullary nails requiring a revision. This clinical case created the motivation for this study. The purpose was to examine the stability of simulated transverse fractures after retrograde flexible intramedullary nail fixation using various nail diameters in synthetic bones. The cross-sectional area of the canal filled by the nails was also determined.
Materials and methods Six pediatric sized synthetic femoral models (2nd generation femur models, Pacific Research Laboratories, Inc., Vashon, WA) were used for mechanical testing. Each composite bone was made of woven fiberglass and epoxy with a solid polyurethane foam core to simulate
0736-0266/$ - see front matter 2005 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.orthres.2005.04.007
1316
J.K. Green et al. / Journal of Orthopaedic Research 23 (2005) 1315–1320
cancellous bone. The femurs measured 28 cm in length and had an oval-shaped mid-shaft intramedullary canal dimension of approximately 8 mm · 11 mm. Synthetic femurs were chosen due to the low availability of cadaveric pediatric femurs. The composite femurs have been shown to be similar to cadaver bone during biomechanical tests with low inter-specimen variability [2,3]. Flexible titanium nails of 2, 3, and 4 mm diameters were used (Synthes Titanium Elastic Nail, Synthes, Oberdorf, Switzerland). The nail manufacturer recommends the use of two equal diameter nails to prevent mal-alignment. To see if different nail diameter combinations would provide greater fracture stability, six combinations of nails were tested: two 2 mm nails (2 + 2 mm), one 2 mm and one 3 mm nail (2 + 3 mm), two 3 mm nails (3 + 3 mm), two 2 mm and one 4 mm nail (2 + 2 + 4 mm), one 3 mm and one 4 mm nail (3 + 4 mm), and two 4 mm nails (4 + 4 mm). Each nail combination was inserted in a random order and tested in each femur. Before mechanical testing, the cross-sectional percent area fill of the femoral canal was determined with a CT scan of one of the intact femurs. Contiguous transverse slices of 1 mm thickness were used. The DICOM images were stacked and reconstructed with 3D medical imaging software (SliceOmatic, Tomovision, Montreal, Canada). Computer models of the six different nail combinations were created within the femoral canal using 3D computer aided design software (Solidworks, Concord, MA). Non-destructive mechanical testing was performed on each intact femur with servo-hydraulic material testing frames (MTS, Eden Prairie, MN). A four-point bend test was done in the sagittal plane to 67 N. The femurs were oriented horizontally with the anterior surface facing upwards (Fig. 1). Each was supported on the posterior surface by rollers spaced 19.9 cm apart and perpendicular to the long axis of the femur. Distally, the roller was just proximal to the condyles in a region that was relatively flat. The proximal roller was just proximal to the lesser trochanter. This roller arrangement prevented specimen rotation during loading due to the wide flat area distally. A vertical load was applied to the anterior surface with two rollers spaced 7 cm apart and centered on either side of the fracture site. The vertical actuator speed was 0.1 mm/s, and the data were collected at 200 Hz. The bending moment was computed from the axial load, axial displacement, and roller placement data. A torsion test was done with the femurs upright (Fig. 2). The proximal and distal femur ends were potted in low melting point alloy. Rectangular fixtures were used, and the alloy captured 2.5 cm of the femurs at each end. Since all the femurs had the same geometry, a single set of alloy molds were made. Each mold was split in the coronal plane so that it could be taken off one specimen and reused on the next. This provided a repeatable means to quickly setup the specimens. The femurs were positioned with the rotational axis of the testing machine going through a point centered between the condyles (mediolaterally and anteroposteriorly) and through the center of the femoral head. Vertical laser levels were used to ensure correct alignment. The distal end of the femur was fixed to a bi-axial load-cell, and the proximal
Fig. 1. Four-point bending setup.
Fig. 2. Torsional test setup.
end fixed to a series of two vertical universal joints attached to the vertical actuator. The rectangular fixtures had shallow holes bored into their plates to accept centering pins located on the load-cell and the distal end of the universal joints. These centering pins ensured that the torsional test setup was identical for each specimen. All femurs were tested in external and internal rotation at a rate of 0.5 deg/s. A 2 N m limit was used. Data were collected at 200 Hz. The last test was a combined axial/bending test with the femurs vertical (Fig. 3). A vertical compressive force was applied at 0.1 mm/s to 85 N to the femoral head through a spherically cut polyethylene cup. The distal end was fixed to a load-cell, and data were collected at 200 Hz. Since the load was applied through the femoral head, a bending moment was applied to the femurs and was not a true axial test of the femoral shaft. Results from a true axial test would be difficult to draw conclusions from since much of the vertical load would be transmitted through the bone and not the implants with a transverse fracture model. By applying the force to the femoral head, the subsequent bending of the structure allowed us to compare the stabilizing effects of each nail combination. After the intact femurs were tested, each was cut mid-shaft to simulate a stable fracture. Since the foam core was solid, a drill bit was used to ream the canal. All of the foam was removed from the diaphysis, but the proximal and distal regions of the femur were essentially left intact to provide support to the nails. Six combinations of nails (2 + 2, 2 + 3, 3 + 3, 2 + 2 + 4, 3 + 4, 4 + 4) were randomly tested in each of the femurs. Nail insertion was performed according to the manufacturerÕs technique guide to have distal nail entry points created on the medial and lateral sides, above the physis. A c-shaped nail layout was used to have the implants fixed distally at the entry point, cross the fracture on the opposite wall of the canal, and fixed proximally in cancellous bone. For a transverse fracture, this provides three points of fixation and allows the nails to share axial loads with the femur. For the unsymmetrical combinations, the larger diameter nail was always inserted on the lateral side. After each femur was instrumented, each nail combinationÕs ability to reduce the fracture and restore proper rotational alignment was noted.
J.K. Green et al. / Journal of Orthopaedic Research 23 (2005) 1315–1320
1317
Data analysis was performed to determine the bending, torsional, and axial/bending stiffnesses. For the four-point bend, the stiffness was calculated over a force range of 40–65 N. In torsion for the intact femurs, the range was 500–2000 N mm. The instrumented femurs would not support as much torque as the intact femurs so the stiffness was instead computed over a 2–10 rotational range. Analysis of the combined axial/bending test was done over 40–85 N. A repeated one-way analysis of variance was used to compare stiffnesses (p < 0.05).
Results
Fig. 3. Combined axial/bending setup. The same three tests performed on the intact femurs were used for the instrumented femurs with the following differences. For the fourpoint bending test, the actuator speed was increased to 1 mm/s (0.1 mm/s for intact) to reduce testing time. The instrumented femurs had considerably less stiffness compared to the intact femurs and required larger amounts of actuator displacement to reach the load limit of 67 N. The higher rate was not used for the intact femurs since relatively small actuator displacements were needed to reach 67 N. The lower rate allowed for better control of the testing machine without overshooting the desired load. For the torsion test a 20 limit was used since the instrumented femurs lacked the torsional stiffness to reach 2 N m at clinically relevant rotational angles.
Based on the CT reconstruction of the composite femur, the available space in the canal for the rods was the least at the mid-shaft, where the 4 + 4 mm nail combination had a 48% canal fill, followed by 34% for the 3 + 4 mm, 32% for the 2 + 2 + 4 mm, 23% for the 3 + 3 mm, 15% for the 2 + 3 mm, and 8% for the 2 + 2 mm. The fracture was reduced by the 2 + 2 mm nail configuration with proper rotational alignment of the bone halves for all six specimens. The 3 + 2 mm, 3 + 3 mm, 4 + 3 mm, and the 2 + 2 + 4 mm nail combinations were each able to reduce the fracture and restore rotational alignment in 83% (5 of 6) of the specimens for each nail combination. The 4 + 4 mm nail configuration was only able to reduce the fracture and restore the rotational alignment in 1 of 6 specimens (17%) due to the greater stiffness of the 4 mm nails and the fact that they occupied a larger volume in the canal, which made insertion and adjustment difficult. When complete reduction did not occur with the 4 + 4 mm nail combination, a posterior gap was present. The intact femurs were significantly stiffer than any of the nail configurations. The average intact bending stiffness was 448 N mm/mm, the torsional stiffness was 2252 N mm/deg for external rotation and 1784 N mm/ deg for internal rotation, and the combined axial/bending stiffness was 983 N/mm.
Fig. 4. Bending stiffness of different nail combinations (mean and standard deviation).
1318
J.K. Green et al. / Journal of Orthopaedic Research 23 (2005) 1315–1320
Fig. 5. Torsional stiffness of different nail combinations (mean and standard deviation).
Fig. 6. Axial/bending stiffness of the intact femur and different nail combinations (mean and standard deviation).
For the four-point bending tests of the instrumented femurs, the 4 + 4 mm rods were significantly stiffer than all other combinations, and the 3 + 4 mm was significantly stiffer than the 2 + 2 mm (Fig. 4). In the torsional tests, the 4 + 4 mm nails were significantly stiffer than all other combinations in external rotation (Fig. 5). In internal rotation, the 4 + 4 mm nails were significantly stiffer than all other combinations, except the 2 + 2 + 4 mm, which in turn was stiffer than the 2 + 2 mm and 3 + 3 mm. The combined axial/bending stiffness for the 3 + 4 mm combination was significantly greater than the 2 + 2 mm, 2 + 2 + 4 mm, and 4 + 4 mm configurations (Fig. 6).
Discussion Flexible titanium nails have been indicated for use in pediatric femoral fractures. These nails permit early weight bearing and reduce some of the mobility issues related to a hip spica cast. Retrograde nails also seem to be less prone to avascular necrosis associated with
reamed intramedullary rods and do not disrupt the epiphyseal plate [4]. The purpose of this study was to compare different combinations of retrograde 2, 3, and 4 mm flexible titanium nails in bending, torsion, and a combined axial/bending test. Unsymmetrical combinations, as well as a three-nail configuration were of particular interest since it may be possible to use several smaller nails in place of one larger nail. Based on the results of the mechanical testing in this study, one might conclude that the 4 mm nails would provide the best torsional stability and bending stiffness. This follows common sense, however, these nails with their greater diameter are stiffer and harder to insert in this size femur and typically did not reduce the fracture (a gap was left posteriorly). Our observations support the nail manufacturerÕs technique guide recommendation of using nails no larger than 40% of the canal width. For this femur, 40% at mid-shaft would be 3.2 mm in the AP view. Therefore, the manufacturer would not recommend the use of two 4 mm nails for this femur. Of interest, for internal rotation, the 2 + 2 + 4 mm combination may be of benefit when larger nails are inap-
J.K. Green et al. / Journal of Orthopaedic Research 23 (2005) 1315–1320
propriate. In this case, even though one 4 mm rod occupies more than 40% of the canal width as seen in an AP X-ray, the mid-shaft cross-sectional percent area fill of the 2 + 2 + 4 mm nail combination is only 32% (compared to 48% for the 4 + 4 mm and 23% for the 3 + 3 mm). Also of interest was the 3 + 4 mm combination. With a percent area fill of 34%, this combination reduced the fracture and restored rotational alignment in 5 out of 6 femurs. In fact, the 3 + 4 mm nail combination had a significantly higher combined axial/bending stiffness compared to all other combinations. This was probably due to the use of the large 4 mm nail on the lateral side and also the fact that the smaller 3 mm nail provided enough room to properly insert the nails and thus reduce the fracture. The two 4 mm nails did not have this extra room and left a gap posteriorly. The torsional stiffness results from this study (25 N mm/deg, 4 mm Ti nails) were similar to published results (48–130 N mm/deg, 3.0–4.0 mm Ti nails) [6,7,9]. Lee et al. [10] found an axial stiffness of 547 N/mm for 3.5 mm stainless steel nails, which compares favorably with our results (688 N/mm, 3 mm Ti nails) even though dissimilar implant materials and a combined axial/bending test rather than a pure axial load were used. The synthetic femurs in these other studies were not the same ones used in this study. Differences in canal width, bone length, and anterior bow may have affected the results. We also aligned our femurs so that the axis of the test machine passed through the femoral head and a point between the condyles for the torsion and axial/bending tests. The other studies aligned the mechanical axis of the femurs co-linear to the test machine axis. Small variations in the alignment of a specimen can produce large changes in the stiffness. Visual observations and manual manipulation of the instrumented femurs revealed that even partial reduction of the transverse fracture allowed axial loads to be shared between the nails and femur. In bending, the larger nail combinations had greater stiffness values, due to their larger cross-sectional area and the fact that they filled the canal near the fracture to a greater extent. The tight fit reduced the amount of implant motion within the femoral canal and produced a stronger construct. The larger nail combinations had the smallest unsupported region since they contacted bone near the fracture. The unsupported region or working length of the implant is measured at the bone-implant contact points on the distal and proximal sides of the fracture [1]. An increase in the unsupported region of the implant results in reduced bending resistance. The smaller implants did contact bone near the fracture, but the fit was loose making the actual unsupported region longer. The stiffness values for the smaller nails might have been larger if they did not move within the canal. Widening of the fracture was also observed for the 2 + 2 mm nails. In torsion, larger nail diameters filled more of the canal and
1319
interlocked with each other better, further resisting movement compared to the small nails. Again, the smaller nails had movement within the canal and in torsion were limited by proximal end fixation in the simulated cancellous bone. Our study has limitations. First, composite femurs were used, and the foam may have behaved differently than cancellous bone in the proximal femur region where the ends of the nails were anchored. The second-generation synthetic femurs used have a relatively stiff foam core, and proximal anchoring of the nails was difficult. Repeated insertion of the different nail combinations may have also deteriorated the foam although care was taken to ensure that each nail was seated as firmly as possible. Friction may also be a limitation. The insertion of the nails was more difficult in the synthetic femurs than during in vivo procedures even in the mid-shaft region. This may be due to small amounts of residual foam in the diaphysis of the synthetic femurs and the absence of bone marrow or moisture to act as lubricant. Because of these factors, the friction between the synthetic femurs and nails may have affected the results. However, these synthetic femurs were nearly identical to each other in terms of shape and composition and allowed the different nail sizes to be compared directly, as opposed to the use of cadaver bones where large inter-specimen variability often affects biomechanical tests due to differences in material properties and geometry. Another limitation is that the study looked only at transverse fractures and did not consider spiral, comminuted, or butterfly fractures. Also, the nail manufacturer does not recommend the use of more than two nails in a femur or unsymmetrical nail combinations to treat femoral fractures. This study was based on single cycle tests and would require further cyclical testing of unsymmetrical and three-nail fixation before it can be recommended clinically. This study has provided mechanical test data showing that several different nail configurations can be used to reduce a transverse femoral fracture. Although a pair of large nails (single nail diameters >40% of the canal width) provides the most torsional and bending stiffness, they were the worst at reducing a fracture gap and should not be used if the fracture cannot be properly reduced. A large nail in combination with one or two smaller nails may provide better stabilization and healing. In all cases, the fractured femur should be protected from high torsional and bending loads.
Acknowledgements AO Synthes donated the flexible nails and surgical instruments. No financial support was received for this study.
1320
J.K. Green et al. / Journal of Orthopaedic Research 23 (2005) 1315–1320
References [1] Browner BD, Jupiter JB, Levine AM, Trafton PG. 2nd ed. Skeletal trauma: fractures, dislocations, ligamentous injuries, vol. 1. Philadelphia: Saunders; 1998. [2] Cristolofini L, Viceconti M. Mechanical validation of whole bone composite tibia models. J Biomech 2000;33:279–88. [3] Cristolofini L, Viceconti M, Cappello A, Toni A. Mechanical validation of whole bone composite femur models. J Biomech 1996;29:525–35. [4] Flynn JM, Hresko T, Reynolds RAK, et al. Titanium elastic nails for pediatric femur fractures: a multicenter study of early results with analysis of complications. J Pediatr Orthoped 2001;21:4–8. [5] Flynn JM, Luedtke LM, Ganley TJ, et al. Comparison of titanium elastic nails with traction and a spica cast to treat femoral fractures in children. J Bone Joint Surg 2004;86-A:770–7. [6] Fricka KB, Mahar AT, Lee SS, Newton PO. Biomechanical analysis of antegrade and retrograde flexible intramedullary nail fixation of pediatric femoral fractures using a synthetic bone model. J Pediatr Orthopaed 2004;24:167–71.
[7] Gwyn DT, Olney BW, Dart BR, Czuwala PJ. Rotational control of various pediatric femur fractures stabilized with titanium elastic intramedullary nails. J Pediatr Orthopaed 2004;24: 172–7. [8] Heinrich SD, Drvaric DM, Darr K, MacEwen GD. The operative stabilization of pediatric diaphyseal femur fractures with flexible intramedullary nails: a prospective analysis. J Pediatr Orthopaed 1994;14:501–7. [9] Kiely N. Mechanical properties of different combinations of flexible nails in a model of a pediatric femoral fracture. J Pediatr Orthopaed 2002;22:424–7. [10] Lee SS, Mahar AT, Newton PO. Ender nail fixation of pediatric femur fractures: a biomechanical analysis. J Pediatr Orthopaed 2001;21:442–5. [11] Ligier JN, Metaizeau JP, Prevot J, Lascombes P. Elastic stable intramedullary nailing of femoral shaft fractures in children. J Bone Joint Surg 1988;70-B:74–7. [12] Mazda K, Khairouni A, Pennecot GF, Bensahel H. Closed flexible intramedullary nailing of the femoral shaft fractures in children. J Pediatr Orthopaed Part B 1997;6:198–202.
A biomechanical study on flexible intramedullary nails used to treat pediatric femoral fractures Jason K. Green *, Frederick W. Werner, Raman Dhawan, Peter J. Evans, Sean Kelley, Dwight A. Webster Department of Orthopedic Surgery, SUNY Upstate Medical University, 3214 Institute for Human Performance, 505 Irving Avenue Syracuse, NY 13210, USA Accepted 21 April 2005
Abstract Flexible intramedullary nails have been indicated to treat femoral fractures in pediatric patients. The purpose of this study was to examine the stability of simulated transverse fractures after retrograde intramedullary flexible nail fixation. Various nail diameter combinations were tested using composite femurs in bending, torsion, and a combined axial/bending test where a vertical compressive force was applied to the femoral head. The cross-sectional percent area fill of the nails within the femurs was also determined. In 4 point bending, the greatest repair stiffness was 12% of the intact stiffness. In torsion, the greatest stiffness was 1% of the intact stiffness for either internal or external rotation. The greatest repair stiffness was 80% of the intact stiffness for a compressive load applied to the femoral head. Nail combinations with single nail diameters greater than 40% of the mid-shaft canal width, as measured from an AP radiograph, prevented the fracture from being reduced and left a posterior gap. Flexible intramedullary nails may be of value in the treatment of pediatric femoral fractures, but care must be taken to insert nails that are correctly sized for the canal and to protect the healing fracture from high torsional and bending loads. 2005 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. Keywords: Pediatric femoral fractures; Flexible intramedullary nail
Introduction Flexible intramedullary nails have been successfully used in treating pediatric femur fractures [4,5,8,11,12]. These have been used in an age group where conservative treatment in a hip spica cast can be difficult on the patient and family due to limited mobility. The reduced immobilization time associated with flexible intramedullary nails is also believed to aid the healing process and eliminate complications of external fixators [4,5]. The optimal diameter and number of nails that should be introduced into the canal to maximize stability is unknown. Although the literature shows that these *
Corresponding author. Tel.: +1 315 464 5228; fax: +1 315 464 6638. E-mail address: [email protected] (J.K. Green).
implants have been successful, a clinical complication was observed by the authors in which a pediatric patient grossly bent a pair of retrograde flexible intramedullary nails requiring a revision. This clinical case created the motivation for this study. The purpose was to examine the stability of simulated transverse fractures after retrograde flexible intramedullary nail fixation using various nail diameters in synthetic bones. The cross-sectional area of the canal filled by the nails was also determined.
Materials and methods Six pediatric sized synthetic femoral models (2nd generation femur models, Pacific Research Laboratories, Inc., Vashon, WA) were used for mechanical testing. Each composite bone was made of woven fiberglass and epoxy with a solid polyurethane foam core to simulate
0736-0266/$ - see front matter 2005 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.orthres.2005.04.007
1316
J.K. Green et al. / Journal of Orthopaedic Research 23 (2005) 1315–1320
cancellous bone. The femurs measured 28 cm in length and had an oval-shaped mid-shaft intramedullary canal dimension of approximately 8 mm · 11 mm. Synthetic femurs were chosen due to the low availability of cadaveric pediatric femurs. The composite femurs have been shown to be similar to cadaver bone during biomechanical tests with low inter-specimen variability [2,3]. Flexible titanium nails of 2, 3, and 4 mm diameters were used (Synthes Titanium Elastic Nail, Synthes, Oberdorf, Switzerland). The nail manufacturer recommends the use of two equal diameter nails to prevent mal-alignment. To see if different nail diameter combinations would provide greater fracture stability, six combinations of nails were tested: two 2 mm nails (2 + 2 mm), one 2 mm and one 3 mm nail (2 + 3 mm), two 3 mm nails (3 + 3 mm), two 2 mm and one 4 mm nail (2 + 2 + 4 mm), one 3 mm and one 4 mm nail (3 + 4 mm), and two 4 mm nails (4 + 4 mm). Each nail combination was inserted in a random order and tested in each femur. Before mechanical testing, the cross-sectional percent area fill of the femoral canal was determined with a CT scan of one of the intact femurs. Contiguous transverse slices of 1 mm thickness were used. The DICOM images were stacked and reconstructed with 3D medical imaging software (SliceOmatic, Tomovision, Montreal, Canada). Computer models of the six different nail combinations were created within the femoral canal using 3D computer aided design software (Solidworks, Concord, MA). Non-destructive mechanical testing was performed on each intact femur with servo-hydraulic material testing frames (MTS, Eden Prairie, MN). A four-point bend test was done in the sagittal plane to 67 N. The femurs were oriented horizontally with the anterior surface facing upwards (Fig. 1). Each was supported on the posterior surface by rollers spaced 19.9 cm apart and perpendicular to the long axis of the femur. Distally, the roller was just proximal to the condyles in a region that was relatively flat. The proximal roller was just proximal to the lesser trochanter. This roller arrangement prevented specimen rotation during loading due to the wide flat area distally. A vertical load was applied to the anterior surface with two rollers spaced 7 cm apart and centered on either side of the fracture site. The vertical actuator speed was 0.1 mm/s, and the data were collected at 200 Hz. The bending moment was computed from the axial load, axial displacement, and roller placement data. A torsion test was done with the femurs upright (Fig. 2). The proximal and distal femur ends were potted in low melting point alloy. Rectangular fixtures were used, and the alloy captured 2.5 cm of the femurs at each end. Since all the femurs had the same geometry, a single set of alloy molds were made. Each mold was split in the coronal plane so that it could be taken off one specimen and reused on the next. This provided a repeatable means to quickly setup the specimens. The femurs were positioned with the rotational axis of the testing machine going through a point centered between the condyles (mediolaterally and anteroposteriorly) and through the center of the femoral head. Vertical laser levels were used to ensure correct alignment. The distal end of the femur was fixed to a bi-axial load-cell, and the proximal
Fig. 1. Four-point bending setup.
Fig. 2. Torsional test setup.
end fixed to a series of two vertical universal joints attached to the vertical actuator. The rectangular fixtures had shallow holes bored into their plates to accept centering pins located on the load-cell and the distal end of the universal joints. These centering pins ensured that the torsional test setup was identical for each specimen. All femurs were tested in external and internal rotation at a rate of 0.5 deg/s. A 2 N m limit was used. Data were collected at 200 Hz. The last test was a combined axial/bending test with the femurs vertical (Fig. 3). A vertical compressive force was applied at 0.1 mm/s to 85 N to the femoral head through a spherically cut polyethylene cup. The distal end was fixed to a load-cell, and data were collected at 200 Hz. Since the load was applied through the femoral head, a bending moment was applied to the femurs and was not a true axial test of the femoral shaft. Results from a true axial test would be difficult to draw conclusions from since much of the vertical load would be transmitted through the bone and not the implants with a transverse fracture model. By applying the force to the femoral head, the subsequent bending of the structure allowed us to compare the stabilizing effects of each nail combination. After the intact femurs were tested, each was cut mid-shaft to simulate a stable fracture. Since the foam core was solid, a drill bit was used to ream the canal. All of the foam was removed from the diaphysis, but the proximal and distal regions of the femur were essentially left intact to provide support to the nails. Six combinations of nails (2 + 2, 2 + 3, 3 + 3, 2 + 2 + 4, 3 + 4, 4 + 4) were randomly tested in each of the femurs. Nail insertion was performed according to the manufacturerÕs technique guide to have distal nail entry points created on the medial and lateral sides, above the physis. A c-shaped nail layout was used to have the implants fixed distally at the entry point, cross the fracture on the opposite wall of the canal, and fixed proximally in cancellous bone. For a transverse fracture, this provides three points of fixation and allows the nails to share axial loads with the femur. For the unsymmetrical combinations, the larger diameter nail was always inserted on the lateral side. After each femur was instrumented, each nail combinationÕs ability to reduce the fracture and restore proper rotational alignment was noted.
J.K. Green et al. / Journal of Orthopaedic Research 23 (2005) 1315–1320
1317
Data analysis was performed to determine the bending, torsional, and axial/bending stiffnesses. For the four-point bend, the stiffness was calculated over a force range of 40–65 N. In torsion for the intact femurs, the range was 500–2000 N mm. The instrumented femurs would not support as much torque as the intact femurs so the stiffness was instead computed over a 2–10 rotational range. Analysis of the combined axial/bending test was done over 40–85 N. A repeated one-way analysis of variance was used to compare stiffnesses (p < 0.05).
Results
Fig. 3. Combined axial/bending setup. The same three tests performed on the intact femurs were used for the instrumented femurs with the following differences. For the fourpoint bending test, the actuator speed was increased to 1 mm/s (0.1 mm/s for intact) to reduce testing time. The instrumented femurs had considerably less stiffness compared to the intact femurs and required larger amounts of actuator displacement to reach the load limit of 67 N. The higher rate was not used for the intact femurs since relatively small actuator displacements were needed to reach 67 N. The lower rate allowed for better control of the testing machine without overshooting the desired load. For the torsion test a 20 limit was used since the instrumented femurs lacked the torsional stiffness to reach 2 N m at clinically relevant rotational angles.
Based on the CT reconstruction of the composite femur, the available space in the canal for the rods was the least at the mid-shaft, where the 4 + 4 mm nail combination had a 48% canal fill, followed by 34% for the 3 + 4 mm, 32% for the 2 + 2 + 4 mm, 23% for the 3 + 3 mm, 15% for the 2 + 3 mm, and 8% for the 2 + 2 mm. The fracture was reduced by the 2 + 2 mm nail configuration with proper rotational alignment of the bone halves for all six specimens. The 3 + 2 mm, 3 + 3 mm, 4 + 3 mm, and the 2 + 2 + 4 mm nail combinations were each able to reduce the fracture and restore rotational alignment in 83% (5 of 6) of the specimens for each nail combination. The 4 + 4 mm nail configuration was only able to reduce the fracture and restore the rotational alignment in 1 of 6 specimens (17%) due to the greater stiffness of the 4 mm nails and the fact that they occupied a larger volume in the canal, which made insertion and adjustment difficult. When complete reduction did not occur with the 4 + 4 mm nail combination, a posterior gap was present. The intact femurs were significantly stiffer than any of the nail configurations. The average intact bending stiffness was 448 N mm/mm, the torsional stiffness was 2252 N mm/deg for external rotation and 1784 N mm/ deg for internal rotation, and the combined axial/bending stiffness was 983 N/mm.
Fig. 4. Bending stiffness of different nail combinations (mean and standard deviation).
1318
J.K. Green et al. / Journal of Orthopaedic Research 23 (2005) 1315–1320
Fig. 5. Torsional stiffness of different nail combinations (mean and standard deviation).
Fig. 6. Axial/bending stiffness of the intact femur and different nail combinations (mean and standard deviation).
For the four-point bending tests of the instrumented femurs, the 4 + 4 mm rods were significantly stiffer than all other combinations, and the 3 + 4 mm was significantly stiffer than the 2 + 2 mm (Fig. 4). In the torsional tests, the 4 + 4 mm nails were significantly stiffer than all other combinations in external rotation (Fig. 5). In internal rotation, the 4 + 4 mm nails were significantly stiffer than all other combinations, except the 2 + 2 + 4 mm, which in turn was stiffer than the 2 + 2 mm and 3 + 3 mm. The combined axial/bending stiffness for the 3 + 4 mm combination was significantly greater than the 2 + 2 mm, 2 + 2 + 4 mm, and 4 + 4 mm configurations (Fig. 6).
Discussion Flexible titanium nails have been indicated for use in pediatric femoral fractures. These nails permit early weight bearing and reduce some of the mobility issues related to a hip spica cast. Retrograde nails also seem to be less prone to avascular necrosis associated with
reamed intramedullary rods and do not disrupt the epiphyseal plate [4]. The purpose of this study was to compare different combinations of retrograde 2, 3, and 4 mm flexible titanium nails in bending, torsion, and a combined axial/bending test. Unsymmetrical combinations, as well as a three-nail configuration were of particular interest since it may be possible to use several smaller nails in place of one larger nail. Based on the results of the mechanical testing in this study, one might conclude that the 4 mm nails would provide the best torsional stability and bending stiffness. This follows common sense, however, these nails with their greater diameter are stiffer and harder to insert in this size femur and typically did not reduce the fracture (a gap was left posteriorly). Our observations support the nail manufacturerÕs technique guide recommendation of using nails no larger than 40% of the canal width. For this femur, 40% at mid-shaft would be 3.2 mm in the AP view. Therefore, the manufacturer would not recommend the use of two 4 mm nails for this femur. Of interest, for internal rotation, the 2 + 2 + 4 mm combination may be of benefit when larger nails are inap-
J.K. Green et al. / Journal of Orthopaedic Research 23 (2005) 1315–1320
propriate. In this case, even though one 4 mm rod occupies more than 40% of the canal width as seen in an AP X-ray, the mid-shaft cross-sectional percent area fill of the 2 + 2 + 4 mm nail combination is only 32% (compared to 48% for the 4 + 4 mm and 23% for the 3 + 3 mm). Also of interest was the 3 + 4 mm combination. With a percent area fill of 34%, this combination reduced the fracture and restored rotational alignment in 5 out of 6 femurs. In fact, the 3 + 4 mm nail combination had a significantly higher combined axial/bending stiffness compared to all other combinations. This was probably due to the use of the large 4 mm nail on the lateral side and also the fact that the smaller 3 mm nail provided enough room to properly insert the nails and thus reduce the fracture. The two 4 mm nails did not have this extra room and left a gap posteriorly. The torsional stiffness results from this study (25 N mm/deg, 4 mm Ti nails) were similar to published results (48–130 N mm/deg, 3.0–4.0 mm Ti nails) [6,7,9]. Lee et al. [10] found an axial stiffness of 547 N/mm for 3.5 mm stainless steel nails, which compares favorably with our results (688 N/mm, 3 mm Ti nails) even though dissimilar implant materials and a combined axial/bending test rather than a pure axial load were used. The synthetic femurs in these other studies were not the same ones used in this study. Differences in canal width, bone length, and anterior bow may have affected the results. We also aligned our femurs so that the axis of the test machine passed through the femoral head and a point between the condyles for the torsion and axial/bending tests. The other studies aligned the mechanical axis of the femurs co-linear to the test machine axis. Small variations in the alignment of a specimen can produce large changes in the stiffness. Visual observations and manual manipulation of the instrumented femurs revealed that even partial reduction of the transverse fracture allowed axial loads to be shared between the nails and femur. In bending, the larger nail combinations had greater stiffness values, due to their larger cross-sectional area and the fact that they filled the canal near the fracture to a greater extent. The tight fit reduced the amount of implant motion within the femoral canal and produced a stronger construct. The larger nail combinations had the smallest unsupported region since they contacted bone near the fracture. The unsupported region or working length of the implant is measured at the bone-implant contact points on the distal and proximal sides of the fracture [1]. An increase in the unsupported region of the implant results in reduced bending resistance. The smaller implants did contact bone near the fracture, but the fit was loose making the actual unsupported region longer. The stiffness values for the smaller nails might have been larger if they did not move within the canal. Widening of the fracture was also observed for the 2 + 2 mm nails. In torsion, larger nail diameters filled more of the canal and
1319
interlocked with each other better, further resisting movement compared to the small nails. Again, the smaller nails had movement within the canal and in torsion were limited by proximal end fixation in the simulated cancellous bone. Our study has limitations. First, composite femurs were used, and the foam may have behaved differently than cancellous bone in the proximal femur region where the ends of the nails were anchored. The second-generation synthetic femurs used have a relatively stiff foam core, and proximal anchoring of the nails was difficult. Repeated insertion of the different nail combinations may have also deteriorated the foam although care was taken to ensure that each nail was seated as firmly as possible. Friction may also be a limitation. The insertion of the nails was more difficult in the synthetic femurs than during in vivo procedures even in the mid-shaft region. This may be due to small amounts of residual foam in the diaphysis of the synthetic femurs and the absence of bone marrow or moisture to act as lubricant. Because of these factors, the friction between the synthetic femurs and nails may have affected the results. However, these synthetic femurs were nearly identical to each other in terms of shape and composition and allowed the different nail sizes to be compared directly, as opposed to the use of cadaver bones where large inter-specimen variability often affects biomechanical tests due to differences in material properties and geometry. Another limitation is that the study looked only at transverse fractures and did not consider spiral, comminuted, or butterfly fractures. Also, the nail manufacturer does not recommend the use of more than two nails in a femur or unsymmetrical nail combinations to treat femoral fractures. This study was based on single cycle tests and would require further cyclical testing of unsymmetrical and three-nail fixation before it can be recommended clinically. This study has provided mechanical test data showing that several different nail configurations can be used to reduce a transverse femoral fracture. Although a pair of large nails (single nail diameters >40% of the canal width) provides the most torsional and bending stiffness, they were the worst at reducing a fracture gap and should not be used if the fracture cannot be properly reduced. A large nail in combination with one or two smaller nails may provide better stabilization and healing. In all cases, the fractured femur should be protected from high torsional and bending loads.
Acknowledgements AO Synthes donated the flexible nails and surgical instruments. No financial support was received for this study.
1320
J.K. Green et al. / Journal of Orthopaedic Research 23 (2005) 1315–1320
References [1] Browner BD, Jupiter JB, Levine AM, Trafton PG. 2nd ed. Skeletal trauma: fractures, dislocations, ligamentous injuries, vol. 1. Philadelphia: Saunders; 1998. [2] Cristolofini L, Viceconti M. Mechanical validation of whole bone composite tibia models. J Biomech 2000;33:279–88. [3] Cristolofini L, Viceconti M, Cappello A, Toni A. Mechanical validation of whole bone composite femur models. J Biomech 1996;29:525–35. [4] Flynn JM, Hresko T, Reynolds RAK, et al. Titanium elastic nails for pediatric femur fractures: a multicenter study of early results with analysis of complications. J Pediatr Orthoped 2001;21:4–8. [5] Flynn JM, Luedtke LM, Ganley TJ, et al. Comparison of titanium elastic nails with traction and a spica cast to treat femoral fractures in children. J Bone Joint Surg 2004;86-A:770–7. [6] Fricka KB, Mahar AT, Lee SS, Newton PO. Biomechanical analysis of antegrade and retrograde flexible intramedullary nail fixation of pediatric femoral fractures using a synthetic bone model. J Pediatr Orthopaed 2004;24:167–71.
[7] Gwyn DT, Olney BW, Dart BR, Czuwala PJ. Rotational control of various pediatric femur fractures stabilized with titanium elastic intramedullary nails. J Pediatr Orthopaed 2004;24: 172–7. [8] Heinrich SD, Drvaric DM, Darr K, MacEwen GD. The operative stabilization of pediatric diaphyseal femur fractures with flexible intramedullary nails: a prospective analysis. J Pediatr Orthopaed 1994;14:501–7. [9] Kiely N. Mechanical properties of different combinations of flexible nails in a model of a pediatric femoral fracture. J Pediatr Orthopaed 2002;22:424–7. [10] Lee SS, Mahar AT, Newton PO. Ender nail fixation of pediatric femur fractures: a biomechanical analysis. J Pediatr Orthopaed 2001;21:442–5. [11] Ligier JN, Metaizeau JP, Prevot J, Lascombes P. Elastic stable intramedullary nailing of femoral shaft fractures in children. J Bone Joint Surg 1988;70-B:74–7. [12] Mazda K, Khairouni A, Pennecot GF, Bensahel H. Closed flexible intramedullary nailing of the femoral shaft fractures in children. J Pediatr Orthopaed Part B 1997;6:198–202.