Effects of plate location and selection on the stability of midshaft clavicle osteotomies: A biomechanical study M. R. Iannotti, MD, L. A. Crosby, MD, P. Stafford, MD, Greg Grayson, MSE, and R. Goulet, PhD, Chattanooga, Tenn
Operative fixation of midshaft clavicle fractures is controversial with few biomechanical data to assist surgical decision making. The purpose of this 2-phase biomechanical investigation is to report on the effects of plate location and selection on the stability of midshaft clavicle fractures. Thirty matched pairs of human adult formalin-fixed clavicles were used. In the first phase, in which a 3.5-mm reconstruction plate and simulated midshaft transverse clavicle osteotomies were used, we observed the effect of superior plate placement compared with anterior placement on fracture rigidity, construct stiffness, and strength. In the second phase, in which simulated midshaft oblique clavicle osteotomies were repaired on the superior aspect, we compared the fracture rigidity, construct stiffness, and strength of the 3.5-mm reconstruction, 3.5-mm limited contact dynamic compression (LCDC), and 2.7-mm dynamic compression (DC) plates. Intact clavicles were prepared, potted, and tested for axial and torsional stiffness in an Instron test frame equipped with gimbaled fixtures. Clavicles were band-sawed to simulate an osteotomy, repaired, re-mounted on the test frame with shear and opening extensometers placed across the osteotomy site, and then tested to observe axial and torsional fracture rigidity and stiffness. Constructs were then loaded to failure in compression. First-order regressions were used to estimate fracture rigidity (in kilonewtons per millimeter) and retained construct stiffness (in kilonewtons per millimeter), whereas the maximum applied compressive load at collapse or gross deformation determined the failure load. Values for the From Biomechanics and Materials Testing Laboratory, University of Tennessee Chattanooga College of Engineering and Computer Science, Chattanooga, Tenn. Funding for this study was provided by the University of Tennessee Center for Excellence in Computer Applications in support of the UTC Biomechanics and Material Testing Laboratory, Chattanooga, Tenn. Reprint requests: M. R. Iannotti, MD, 9100 E Florida Ave, #22307, Denver, CO 80231. Copyright © 2002 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/2002/$35.00 ⫹ 0 32/1/125805 doi:10.1067/mse.2002.125805
comparison groups were tested for significance at the 95% confidence level. In the first phase we found that constructs plated at the superior aspect of the clavicle exhibited significantly greater fracture rigidity and mean retained stiffness than the anterior location (P ⬍ .05). In the second phase we found that the torsional fracture rigidity of LCDC-plated constructs significantly exceeded that of the reconstruction and DC plates (P ⬍ .05), whereas the axial fracture rigidity of the LCDC-plated constructs significantly exceeded that of the reconstruction plate (P ⬍ .05). In retained stiffness the performance of the LCDC-plated constructs significantly exceeded that of the DC plate in torsion (P ⬍ .05), whereas in load to failure the LCDC plate withstood significantly more compressive load than the reconstruction plate (P ⬍ .05). We concluded that clavicles plated at the superior aspect exhibit significantly greater biomechanical stability than those plated at the anterior aspect. Furthermore, we concluded that the LCDC plate offers significantly greater biomechanical stability than the reconstruction and DC plates. (J Shoulder Elbow Surg 2002;11:457-62.)
INTRODUCTION Most clavicle fractures have been treated nonoperatively with acceptable clinical outcomes.12 Operative treatment may be indicated for the following: neurovascular injury, open injury, tented or necrotic skin, and irreducible, 100% displaced midshaft fractures.1,13,15 Follow-up studies have demonstrated that acute displacement of 17 to 20 mm or more in adult midshaft clavicle fractures is associated with a high rate of malunion and nonunion.7,9 Long-term clinical results have demonstrated that 20 mm or more of shortening is associated with nonunion in up to 15% of fractures, thoracic outlet syndrome in 30%, and dissatisfaction in 30% of patients.7 Many different methods of operative fixation of midshaft clavicle fractures have been recommended, including smooth pins, threaded wires, and plate fixation.4,11 Their associated complications include smooth pin migration, threaded wire rotational insta-
457
458
Iannotti et al
J Shoulder Elbow Surg September/October 2002
Figure 1 Photograph from phase II illustrating a clavicle specimen, stabilized with the 2.7-mm DC place, in the gimbaled fixture. The extensometer is positioned across the osteotomy site.
bility, plate breakage, and plate removal because of discomfort.1,2,4 Early attempts at operative intervention lacked biomechanical data for implant location and selection. The one biomechanical report available on midshaft clavicle osteotomy fixation demonstrated that superior placement of a one-third tubular plate was more stable than anterior placement in clavicles that had transverse midshaft osteotomies and an intact inferior cortex.5 One-third tubular plates have not been recommended in the treatment of clavicle fractures because of high rates of plate fatigue and failure in the clinical setting. The biomechanical data discovered in this experiment may allow orthopaedic surgeons to make more definitive clinical decisions about the location and selection of plates for midshaft clavicle fracture treatment. The purpose of this study is to report on the effects of plate location and selection on the stability of midshaft clavicle osteotomies. MATERIALS AND METHODS Thirty matched pairs of human adult formalin-fixed clavicles were stripped of soft tissue, underwent radiography, were inspected, and were found to be free of structural deficiency. The mean age for all specimens was 65 years (range, 55-75 years). In the first phase of the study the effect of location on construct performance was determined by use of 6 matched pairs through comparisons of construct stiffness and fracture rigidity with the right and left clavicles assigned the anterior and superior locations, respectively, for a 3.5-mm reconstruction plate. In the second phase the effect of 3 plates on biomechanical performance was determined by use of 24 clavicle pairs divided into 3 groups,
representing reconstruction, limited contact dynamic compression (LCDC), and dynamic compression (DC) plates. Load tests were conducted in axial compression mode and torsion rotational mode on an Instron 8521S test frame (Canton, Mass) equipped with 2 gimbaled fixtures. Figure 1 illustrates a clavicle cemented at each end into a concentric 2-ring set capable of transmitting axial torsion. Each ring may rotate about its axis, thus eliminating bending moment constraints at each end. Axial compression and torsion rotational modes used the setup shown in Figure 1. Applied load was measured through use of the Instron 11048 bi-axial load cell, and construct deflections for stiffness tests were measured through use of Instron linear and rotary variable differential transformer position transducers. Motion at the fracture site was measured with an Instron 2620-825 extensometer. Normal and shear knife-edge attachments were used for axial compression and torsion rotational mode measurements, respectively. Instron MAX programmable control and data acquisition was used to apply predetermined load types and magnitudes automatically and to record load, position, displacement, and time data at 10 Hz. Testing was conducted separately in 2 cyclical modes, axial compression and then torsion rotational, in order to obtain in vitro data that may resemble in vivo clavicle biomechanics. Axial tests were conducted in load control, whereas torsion tests were conducted in rotary position control. The Instron unit could not perform torsion rotational testing in torque control. Control waveforms for both modes were saw-toothed in shape, with saw-toothed segments alternating above and below the x-axis. Alternating segments about the x-axis represented relative compression and tension in axial testing and internal and external rotation in torsion rotational testing. The magnitude of each subsequent segment was 10% greater than the previous segment’s magnitude. The maximum amplitudes in the axial compression and torsion rotational testing after 10 cycles
Iannotti et al
J Shoulder Elbow Surg Volume 11, Number 5
459
Table I Comparison of the physical characteristics of the 3 plates Characteristic Dimensions (length ⫻ width ⫻ height) (mm) Steel type Work type Screw type Screw size Screw no. Cross-sectional moment* Yield strength†
3.5-mm Reconstruction plate 70 ⫻ 10 ⫻ 2.8 316 L stainless steel Annealed Cortical 3.5 mm 6 2 1
3.5-mm LCDC plate 77 ⫻ 11 ⫻ 4 316 L stainless steel Cold-worked Cortical 3.5 mm 6 4 2
2.7-mm DC plate 68 ⫻ 8 ⫻ 2.5 316 L stainless steel Cold-worked Cortical 2.7 mm 8 1 2
*Approximated using reference 3. Approximated by measurement of hardness in this laboratory by first author.
†
were ⫺0.3 ⫾ 0.12 kN and 0.0° ⫾ 7.0°, respectively. The load-to-failure tests were conducted in axial compression via load control through use of the saw-tooth waveform with a progressively increasing segment amplitude until specimen failure occurred. Pilot studies done in this laboratory through trial and error demonstrated that 10 cycles in each mode was sufficient to produce quality data yet avoid complications. Specimens generally sustained 10 cycles in load-to-failure testing; thus, each specimen underwent approximately 50 testing cycles. Most commonly, failure occurred during the compression phase of axial compression testing. Specimens were potted in 3-inch-diameter steel bases with Cerro bismuth alloy (Bellefonte, Pa), which has a melting temperature of 55°C. Each intact specimen was mounted in the test frame for compression and torsion stiffness tests. The intact specimens were then cut with a band saw at the mid shaft to simulate a fracture. Phase I osteotomies were transverse in order to reproduce results from the previous biomechanical study by Harnroongroj and Vanadurongwan.5 Phase II osteotomies were oblique to simulate the most common clinical type of fracture pattern yet noncomminuted to avoid a confounding variable. Clavicles were reduced and plated by standard orthopaedic plating techniques, including compressive bone fixation with clamps and proper drilling, tapping, and screw utilization. Phase II plates, the characteristics of which are summarized in Table I, were placed in the superior position as dictated by phase I results. Each repaired specimen was re-mounted in the test frame with the extensometer positioned to bridge the osteotomy 180° from the plate centerline. After the compression and torsion tests of the plated specimens, the extensometer was removed and the specimen was loaded to failure in compression and then visually inspected by the first author. Recorded time, load, extensometer displacement, and linear variable differential transformer (LVDT) deflection data were analyzed. The extensometers directly measured axial and torsional motion at the fracture site, whereas the LVDT measured axial and torsional motion of the entire construct via the cross-head of the testing frame. With these data, fracture rigidity (in kilonewtons per millimeter) was estimated by first-order regression of applied load as a function of extensometer displacement, and construct stiffness (in kilonewtons per millimeter) was estimated by firstorder regression of applied load as a function of LVDT or construct deflection. The repaired construct stiffness was
normalized with respect to the intact construct stiffness and reported as percent intact stiffness. In this laboratory, rigidity has been used as an accurate and precise reflection of local activity at the osteotomy site and construct stiffness has been used as a more global representation of the entire construct’s biomechanical behavior. Failure load was defined as the maximum applied compressive load at the time of catastrophic failure of the construct, and catastrophic failure was defined as bony fracture or plate failure resulting in the construct’s sudden inability to further withstand applied load. Values for the comparison groups were computed and tested for significance at the 95% confidence level. Although the empirical reality would suggest that the observed rigidity and stiffness are normally distributed, the assumption could not be verified because of the small sample size. Therefore, the reported significance was measured with the nonparametric Wilcoxon test and then verified with the parametric t test for difference of two means and the Bonferroni correction factor test for a 3-way pairwise comparison.14
RESULTS The axial and torsional fracture rigidity and retained stiffness and compressive strength for anteriorly and superiorly plated constructs are shown in Table II. Fracture rigidity of constructs plated superiorly exceeded that of those anteriorly plated by 290% in axial loading and by 58% in torsion (P ⬍ .05). Similarly, the retained stiffness of constructs plated superiorly exceeded that of those plated anteriorly by 41% in axial loading and by 20% in torsion (P ⬍ .05). The axial and torsional fracture rigidity and retained stiffness and compressive strength for reconstruction-, LCDC-, and DC-plated constructs are shown in Table III. The torsional fracture rigidity of LCDCplated constructs exceeded that of the reconstruction and DC plates by 130% and 150%, respectively (P ⬍ .05). The LCDC-plated constructs also exceeded the axial fracture rigidity of the reconstruction plate by 237% (P ⬍ .05). In retained stiffness the performance of the LCDC plate constructs exceeded that of the DC plate by 36% in torsion (P ⬍ .05). In load to failure
460
Iannotti et al
J Shoulder Elbow Surg September/October 2002
Table II Effect of location on fracture rigidity, retained stiffness, and load to failure Axial load Biomechanical property Fracture rigidity Retained stiffness Load to failure
Torsional load
Anterior
Superior
P (␣ ⴝ .05)
Anterior
Superior
P (␣ ⴝ .05)
0.63 kN/mm 73% 1.31 kN
2.50 kN/mm 103% 1.51 kN
⬍.05 ⬍.05 ⬎.20
5.26 kN-m/mm 64% NA
8.33 kN-m/mm 77% NA
⬍.05 ⬍.05 NA
NA, Not available.
Table III Effect of plate type on fracture rigidity, retained stiffness, and load to failure Biomechanical property Reconstruction Fracture rigidity
Intact stiffness
Load to failure
Axial load LCDC
Torsional load DC
1.79 kN/mm 5.74 kN/mm — — 5.74 kN/mm 4.89 kN/mm 1.79 kN/mm — 4.89 kN/mm 69% 61% — — 61% 66% 69% — 66% 0.69 kN 0.94 kN — — 0.94 kN 0.88 kN 0.69 kN — 0.88 kN
P (␣ ⴝ .05) Reconstruction ⬍.05 ⬎.25 ⬍.11 ⬎.25 ⬎.25 ⬎.25 ⬍.05 ⬎.25 ⬍.10
.013 kN-m/mm — .013 kN-m/mm 68% — 68% NA NA NA
LCDC
DC
.030 kN-m/mm — .030 kN-m/mm .012 kN-m/mm — .012 kN-m/mm 79% — 79% 58% — 58% NA NA NA NA NA NA
P (␣ ⴝ .05) ⬍.05 ⬍.05 ⬎.25 ⬍.08 ⬍.05 ⬍.08 NA NA NA
NA, Not available.
the LCDC plate withstood 36% more compressive load than the reconstruction plate (P ⬍ .05). The reconstruction plates failed, with midspan bending deformation accompanied by pullout of the screws closest to the osteotomy site. The LCDC plates did not undergo permanent deformation. The LCDC plate screws pulled out at the holes most distant from the osteotomy site. The DC plates underwent slight deformation with variable pullout in the bone-screw interface. DISCUSSION Prediction of in vivo performance of orthopaedic constructs based on in vitro observation requires careful consideration of many factors, especially the type, direction, and magnitude of applied loading. The functions and kinematics of clavicle motion are well established.6,8,10 Data related to the types, directions, and magnitudes of physiological clavicle forces coincident with this motion were not present in the literature, however. This experiment, therefore, was designed to apply forces of type, direction, and magnitude logically consistent with known clavicle function and motion. The clavicle acts as a strut to enable motion of the arm and, therefore, is subjected to both tensile and compressive loads; thus, the experiment was designed to apply tensile and compressive axial loads. Three-dimensional curvilinear-shaped structures such as the clavicle deflect with torsion or twisting, consistent with the curved beam theory.3 The
experiment was designed to apply torsional loads in position control. In addition, the Instron test frame could not repeatedly perform torsion rotational testing in torque control. The clavicle rotates in 3 planes, and the experiment was designed to apply axial and torsional loads through gimbaled fixtures that simulate the absence of bending constraints at the sternal and acromial ends. In the first phase of this study, which used the 3.5-mm reconstruction plate for its contourability, we compared the biomechanical stability of midshaft clavicle transverse osteotomies plated on the anterior aspect with that on the superior aspect. Our results indicate that plating the superior aspect provides significantly higher fracture rigidity and retained stiffness than anterior plating (P ⬍ .05). We believe that the superior plate is on a tension-bearing surface, and it has a greater moment of resistance because of its larger distance from the inferior cortex compared with the anterior plate. These facts may help explain the higher fracture rigidity and stiffness in the superior position. Phase I results also demonstrated that superiorly and anteriorly plated constructs did not have significantly different strengths (P ⬎ .20). Although this result suggests that there is relatively little effect of location on construct strength, this ambiguous outcome may result from the use of the transverse osteotomy. Free-body mechanical analysis indicates that the reduced transverse clavicle osteotomy is inherently more stable and transfers significantly higher
J Shoulder Elbow Surg Volume 11, Number 5
Iannotti et al
461
Figure 2 Front-view photograph of the 3 plates: 2.7-mm DC (top), 3.5-mm reconstruction (middle), and 3.5-mm LCDC (bottom).
Figure 3 Side profile view of the 3 plates: 2.7-mm DC (top), 3.5-mm reconstruction (middle), and 3.5-mm LCDC (bottom).
normal and shear traction loads than the oblique osteotomy. Construct failure with transverse osteotomies would, therefore, be influenced more by the material compressive and shear strengths of individual cadaveric specimens. Conversely, specimens with oblique osteotomies may be more influenced by the properties of the plate. We chose to use a transverse osteotomy in the first phase of this study to verify the results of a related study by Harnroongroj and Vanadurongwan,5 who reported better performance for superiorly positioned one-third tubular plates used to repair transverse midshaft clavicle osteotomies. In their study the cadaveric specimen was mounted and fixed at the sternal end, while the cantilevered acromial end was loaded in an inferior direction. Our study was designed, however, with two important
differences. We applied a loading configuration more physiologically consistent with known clavicle kinematics and functions and used the reconstruction plate in lieu of the obsolete one-third tubular plate. In the second phase of this study (plating the superior aspect of the clavicle to stabilize midshaft oblique osteotomies), we compared the biomechanical stability of 3 plates: the 3.5-mm reconstruction plate, the 3.5-mm LCDC plate, and the 2.7-mm DC plate (Figures 2 and 3). Our results indicate that the LCDC plate provides a significantly higher axial fracture rigidity than the reconstruction plate (P ⬍ .05), with no difference between LCDC and DC plates (P ⬎ .05). These findings may be attributable to the LCDC’s characteristics, including cold-working, yield strength, and large cross-sectional moment. Load-to-
462
Iannotti et al
J Shoulder Elbow Surg September/October 2002
failure testing confirmed these findings. The annealed reconstruction plate bent in mid span, whereas the cold-worked LCDC plate did not bend and the DC plate bent only slightly. The LCDC plate also provided higher torsional fracture rigidity than the DC and reconstruction plates (P ⬍ .05). These findings may be attributable to the LCDC’s characteristics, including greater width and cross-sectional moment. In summary, the superior performance of the LCDC plate over the reconstruction plate in axial and torsional loading may be attributed to several LCDC plate characteristics, including width, thickness, cold-working, yield strength, and cross-sectional moments. The superior performance of the LCDC plate over the DC plate in torsional fracture rigidity may be attributed to the wider and thicker dimensions of the LCDC plate and its greater torque-resisting moment arm offered by the edge-to-screw centerline dimension. This experiment demonstrated that superior placement of the 3.5-mm LCDC plate for midshaft clavicle osteotomies was the most stable biomechanical model studied. The orthopaedic surgeon may use this information as a guide in clinical decision making regarding plate location and implant selection for midshaft clavicle fractures. Specifically, the 3.5-mm LCDC plate provides superior stabilization and should be placed in the superior anatomic position. We thank Testing and Technologies, Inc, Chattanooga, Tenn, for radiographic services and Colleen Schmitt, MD, MS, for statistical assistance.
REFERENCES
1. Ali Khan MA, Lucas HK. Plating of fractures of the middle third of the clavicle. Injury 1978;9:263-7. 2. Boehme D, Curtis RJ Jr, DeHaan JT, Kay SP, Young DC, Rockwood CA. Non-union of fractures of the mid-shaft of the clavicle. J Bone Joint Surg Am 1991;73:1219-25. 3. Boresi AP, Schmidt RJ, Sidebottom OM. Deflection of curved beams. Advanced mechanics of materials. New York: Wiley & Sons; 1993. p. 385-91. 4. Bostman O, Manninen M, Pihlajamaki H. Complications of plate fixation in fresh displaced midclavicular fractures. J Trauma 1997; 43:778-83. 5. Harnroongroj T, Vanadurongwan V. Biomechanical aspects of plating osteosynthesis of transverse clavicular fracture with and without inferior cortical defect. Clin Biomech 1996;11:290-4. 6. Harrington MA Jr, Keller TS, Seiler JG III, Weikert DR, Moeljanto E, Schwartz HS. Geometric properties and the predicted mechanical behavior of adult human clavicles. J Biomechan 1993;26: 417-26. 7. Hill JM, McGuire MH, Crosby LA. Closed treatment of displaced middle-third fractures of the clavicle gives poor results. J Bone Joint Surg Br 1997;79:537-9. 8. Inman VT, Saunders M. Observations on the function of the clavicle. Calif Med 1946;65:158-66. 9. Jupiter JB, Leffert RD. Non-union of the clavicle. Associated complications and surgical management. J Bone Joint Surg Am 1987; 69:753-60. 10. Ljunggren AE. Clavicular function. Acta Orthop Scand 1979;50: 261-8. 11. Manske DJ, Szabo RM. The operative treatment of mid-shaft clavicular non-unions. J Bone Joint Surg Am 1985;67:1367-71. 12. Neer CS II. Non-union of the clavicle. JAMA 1960;172:1006-11. 13. Poigenfurst J, Rappold G, Fischer W. Plating of fresh clavicular fractures: results of 122 operations. Injury 1992;23:237-41. 14. Scheaffer RL, McClave JT. Comparing two populations: Wilcoxon signed rank test for the paired difference experiment. Statistics for engineers. Boston: Duxbury Press; 1982. p. 372-9. 15. Schwarz N, Hocker K. Osteosynthesis of irreducible fractures of the clavicle with 27-mm ASIF plates. J Trauma 1992;33:179-83.
ANNUAL CONGRESS OF THE FRENCH ARTHROSCOPY SOCIETY The Annual Congress of the French Arthroscopy Society will be held December 5-7, 2002, at the Palais des Congre`s in Paris, France. State of the art sessions, instructional courses, symposiums, oral papers, posters, an international session, and a videoflash session will be offered. There will be a simultaneous translation in English. For more information, please contact the Scientific Secretary, Dr. Laurent Lafosse, Clinique Ge`ne`rale, 4 Chemin Tour de la Reine, 74000 ANNECY; phone: 33 (0) 4 50 33 09 50; fax: 33 (0) 4 50 33 09 77; e-mail:
[email protected]. For information regarding registration, contact Registration, MCO Congre`s, 27 Rue du Four a` Chaux, 13007 MARSEILLE; phone: 33 (0) 4 95 09 38 00; fax: 33 (0) 4 95 09 38 01; e-mail:
[email protected]. For information regarding the organization, contact Organisation, Dr. Philippe Landreau, Clinique Moventis, 36 Bd St Marcel – 75005 PARIS; phone: 33 (0) 1 40 79 40 00; fax: 33 (0) 1 40 79 40 43; e-mail:
[email protected].