- Email: [email protected]
Are conventional reconstruction plates equivalent to precontoured locking plates for distal humerus fracture fixation? A biomechanics cadaver study
Clinical Biomechanics 27 (2012) 697–701
Contents lists available at SciVerse ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
Are conventional reconstruction plates equivalent to precontoured locking plates for distal humerus fracture fixation? A biomechanics cadaver study☆ Ryan C. Koonce a,⁎, Todd H. Baldini b, Steven J. Morgan c a b c
Skagit Regional Clinics, 1400 E. Kincaid Street, Mount Vernon, WA 98274, United States University of Colorado School of Medicine, Department of Orthopaedic Surgery, 13001 E. 17th Place, Aurora, CO 80045, United States Swedish Medical Center, MOTUS Orthopaedics, 701 E Hampden Ave Suite 515, Englewood Co 80113 United States
a r t i c l e
i n f o
Article history: Received 20 November 2011 Accepted 27 March 2012 Keywords: Humerus fracture Fixation Locking plate Biomechanics Cadaver
a b s t r a c t Background: The optimal plate type and configuration for distal humerus fracture fixation has yet to be defined. Available biomechanical studies show conflicting results. No existing studies compare conventional reconstruction plates to newer precontoured distal humerus locking plates in both parallel and perpendicular configurations. Methods: Three groups of humerus specimens were compared via biomechanical testing in a cadaver model simulating metaphyseal comminution. Group 1 consisted of conventional reconstruction plates in a perpendicular configuration. Group 2 used precontoured locking plates in a perpendicular configuration. Group 3 used precontoured locking plates in a parallel configuration. Each group was tested for stiffness in anterior bending, posterior bending, axial compression, and torsion. The specimens then underwent cyclic loading followed by single load to failure in posterior bending. Findings: There was no significant difference between the three groups for anterior bending, posterior bending, axial compression, or torsional stiffness. There was no significant difference in load to failure for any of the three groups. Screw loosening was significantly higher in Group 1 when compared to Groups 2 and 3 after cyclic loading. Interpretation: In the early postoperative period, less expensive perpendicular conventional reconstruction plate constructs provide similar stiffness and load to failure properties to newer precontoured locking plate systems regardless of plate configuration. © 2012 Elsevier Ltd. All rights reserved.
1. Introduction Fractures of the adult distal humerus have been described as grave lesions with poor outcomes since circa 415 BC (Brorson, 2009). Doubleplate osteosynthesis is the current standard for treatment in active adults; however, plate type and configuration are topics of controversy within the literature (Green, 2009; Nauth et al., 2011). Historically, the AO group has recommended treatment with conventional reconstruction plates (CRPs) in a perpendicular configuration (Rüedi et al., 2007); wherein the lateral column plate is placed posteriorly and the medial column plate turned approximately 90° and placed medial to the supracondylar ridge. There is a current trend toward use of precontoured distal humerus locking plates (PDHLPs) in a parallel configuration (Nauth et al., 2011; O'Driscoll, 2005); where plates are placed on the medial and
☆ Location of work: This study was performed at the University of Colorado School of Medicine, Department of Orthopaedic Surgery Biomechanics Laboratory in Aurora, CO, USA. ⁎ Corresponding author. E-mail addresses: [email protected] (R.C. Koonce), [email protected] (T.H. Baldini), [email protected] (S.J. Morgan). 0268-0033/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2012.03.008
lateral columns approximately 180° to each other. These newer plates are attractive as an alternative to CRPs because of angular screw stability and an anatomically precontoured shape, but evidence to recommend the routine use of locking plates over non-locking plates for distal humerus fractures is insufficient (Nauth et al., 2011). Several biomechanical studies with conflicting results have been published comparing CRPs, PDHLPs, and locking compression plates (LCPs) in various configurations. With regard to biomechanical stability of various plate and screw constructs, the following issues are currently unproven: (a) whether newer PDHLPs are superior to CRPs; and (b) whether the parallel plate configuration is superior to the perpendicular configuration. With a lack of consensus on these issues, individual surgeon preference and experience often dictate the choice of implant and implant position for internal fixation of distal humerus fractures (Abzug and Dantuluri, 2010; Schwartz et al., 2006). Currently no single study compares the biomechanical properties of perpendicularly placed CRPs to PDHLPs in both parallel and perpendicular configurations. The aim of our study is to make these comparisons using human cadaver specimens in an established biomechanical testing model (Korner et al., 2003, 2004).
698
R.C. Koonce et al. / Clinical Biomechanics 27 (2012) 697–701
2. Methods 2.1. Specimens Thirty fresh frozen left human upper extremity specimens were obtained (Science Care™, Phoenix, AZ, USA) and stripped of all soft tissue and visually inspected for pathology. The mean age of human cadaver specimens used was 72.3 years (range 51 to 98, standard deviation 10.8). There were 17 humeri from female cadavers and 13 from male cadavers. Total bone mineral density (BMD) of the distal third of each humerus specimen was measured with a dual energy X-ray absorptiometry (DEXA) machine (Hologic, Inc., Bedford, MA, USA). The mean and standard deviation for bone mineral density were mean 0.69 (SD 0.18 g/cm 2). After the DEXA scans were completed the humeri were randomly assigned to three groups of ten. The mean BMDs and standard deviations of the three groups were as follows (g/cm 2): Group 1 (0.69,0.19), Group 2 (0.64,0.22), and Group 3 (0.73,0.12). These were statistically compared using one way ANOVA to ensure there was no significant difference in BMD between the three groups (P = 0.98). The specimens were then wrapped and stored in saline soaked gauze at −20 °C prior to testing. 2.2. Fracture model Specimens were thawed for 12 h at room temperature. First, plates with moldable segments were contoured to fit to the bone. Next, the proximal and distal screw holes of medial and lateral/posterolateral plates were drilled into the bone to ensure an anatomical reduction. A supracondylar distal humerus fracture model was then created using a band saw to cut a transverse 5 mm osteotomy gap just proximal to the olecranon fossa to simulate metaphyseal comminution (OTA/AO type 13-A3.3). The gap was large enough to avoid bone contact between proximal and distal fragments during testing. 2.3. Fracture fixation Three different plating configurations and four plate types were used for fracture fixation. All implants used are Food and Drug Administration approved for distal humerus fracture fixation. Group 1: Non-locking 3.5 mm stainless steel 8-hole (94 mm length) CRPs, (Smith & Nephew, Inc., Memphis, TN, USA), were anatomically mounted around the medial epicondyle on the ulnar column and along the posterolateral surface on the radial column in a perpendicular configuration using 3.5 mm non-locking cortex screws (Fig. 1A, D). Group 2: Plates were placed in a similar manner to plates in Group 1 using 7-hole medial (103 mm) and 7-hole posterolateral (107 mm) stainless steel posterolateral PDHLPs (Smith & Nephew PERI-LOC) using only locking screws in a perpendicular configuration (Fig. 1B, E). Group 3: 7-hole medial (103 mm) and lateral (102 mm) stainless steel PDHLPs (PERI-LOC, Smith & Nephew, Inc., Memphis, TN, USA) were placed on ulnar and radial columns using only locking screws in a parallel configuration (Fig. 1C, F). Plate lengths were chosen to most closely match the overall working distance of the PDHLPs to an 8-hole CRP. Plate thickness is 2.8 mm for all CRPs and 3.1 mm for all PDHLPs. Pilot holes of appropriate size for the screws per the manufacturers' specifications were made with a power drill. Screws were inserted by hand to a tightness of two fingers. A 1.7 Nm torque limiting screw driver was utilized for all locking screw insertions. Proximal to the osteotomy site, three 3.5 mm bicortical screws were placed in every plate. Distal to the osteotomy site, monocortical screws were used to avoid penetration of articular surfaces. CRPs were fixed distally with three monocortical 3.5 mm non-locking screws. PDHLPs were fixed distally with one monocortical 3.5 mm locking screw and three
Fig. 1. Photographs and plain X-rays demonstrating three different testing groups. Group 1 (A and D) = perpendicular CRPs; Group 2 (B and E) = perpendicular PDHLPs; Group 3 (C and F) = parallel PDHLPs.
monocortical 2.7 mm locking screws. Penetration of the olecranon fossa was avoided in all groups. 2.4. Potting the specimens After the humeri were plated they were potted for mechanical testing. The proximal end of the humeri was cut off with a band saw 6 cm from the most proximal end of the plates. Five centimeters of the proximal end of the bone was potted in the center of a 4 in. diameter aluminum tube with Dyna-Cast hard setting urethane (Kindt-Collins Company LLC, Cleveland, OH, USA). Fixing the distal end of the humeri was accomplished by first placing a highly elastic modeling compound (Hasbro, Inc., Pawtucket, RI, USA) over the distal plate and screws. This prevented plates and screws from being rigidly fixed in the potting compound and allowed the screws to back out during testing. The distal end was then potted to within 1 cm of the osteotomy in the center of a 4 in. square tube with Dyna-Cast. 2.5. Mechanical testing Three different types of mechanical testing were performed: stiffness testing, cyclic loading, and single load to failure. All tests were performed on an Instron Model 1321 closed loop servo-hydraulic test machine (Instron Corporation, Canton, MA, USA). All data were collected at 50 Hz on a PC equipped with a Keithley 1802HC (Keithley Instruments, Inc., Cleveland, OH, USA) analog to digital board and TestPoint (Capital Equipment, Corp., Billerica, MA, USA) data acquisition software. Stiffness testing was performed in axial compression (250 N), torsion (+/−1.6 Nm), anterior 4 point bending (4.5 Nm), and posterior 4 point bending (4.5 Nm) (Fig. 2). These load levels were chosen to avoid plastic deformation of the construct while stiffness testing (Korner et al., 2004). Four point bending was used to provide a constant bending moment across the fracture site. All tests were sinusoidal wave forms run for 4 cycles at 0.2 Hz. The stiffness was measured on the 4th cycle between 20% and 80% of the peak applied load.
R.C. Koonce et al. / Clinical Biomechanics 27 (2012) 697–701
699
Fig. 2. Biomechanical testing setup. An Instron Model 1321 closed loop servo-hydraulic test machine (Instron, Corp., Canton, MA, USA) is pictured with a distal humerus specimens following osteosynthesis. The testing setup is shown for (A) axial compression and torsion, and (B) posterior four-point bending.
After the stiffness tests were completed the specimens were cyclically loaded for 4000 cycles at 1 Hz to 4.5 Nm in posterior bending. This frequency and number of cycles is comparable to previous studies, and estimates approximately 3 months of elbow rehabilitation (Korner et al., 2004; Self et al., 1995). All specimens were inspected after cyclic loading for signs of proximal screw loosening, screw or plate fracture, or bone fracture. All specimens were then tested to ultimate load in posterior 4 point bending. Load was applied under displacement control at a rate of 0.1 mm/s until the specimen failed or 5000 N was applied, the limit of the load cell. After ultimate load testing the potting compound on the distal end of each humerus was carefully removed and signs of failure were recorded. Failure was recorded as implant breakage, bone fracture, or plastic deformation. Distal screw loosening was noted by visually inspecting for screw backout (Fig. 3).
2.6. Statistical analysis A prospective power analysis was performed to determine the required sample size (n) with α = 0.05 and power = 0.8 using JMP data analysis software (SAS Institute Inc., Cary, NC, USA). The power analysis determined an n of 10 was sufficient to obtain statistical significance. After testing, a one way analysis of variance (ANOVA) was used to determine if there was a statistical difference (P b 0.05) in recorded values between the groups assuming the data is normally distributed. A Pearson Chi Square test was used to determine if
there were a significantly greater number of loose screws between groups. 3. Results Using one-way ANOVA, there was no significant difference between the three groups for axial (P = 0.33), torsion (P = 0.48), anterior bending (P = 0.15), or posterior bending (P = 0.35) stiffness. After cyclic testing, specimens were loaded to 5000 N in posterior bending. All 30 failed in plastic deformation. No specimens failed by implant breakage or humerus fracture. Load to failure was defined as the elastic limit for each specimen using standard load–displacement curves. There was no statistically significant difference for load to failure between the three groups (P = 0.55). Results for stiffness testing and load to failure are summarized in Table 1. No specimens demonstrated screw loosening in the diaphysis. All 10 of the Group 1 posterolateral CRPs had two distal screws loose and 8 had all three distal screws loose. Fig. 3 shows an example of a loose screw in Group 1. Four of the Group 1 medial CRPs had at least one loose distal screw. Group 2 had one medial plate with a distal screw loose and one posterior–lateral plate with a distal screw loose. Group 3 had only one distal screw loose on a medial plate. A summary of total screw loosening is shown in Fig. 4. Using a Pearson Chi Square test revealed a statistically significant association between screw type (locking vs. non-locking) and screw loosening. The CRP group, using only non-locking screws, was significantly more likely to have loose screws than the PDHLP groups which used only locking screws (P b 0.001). 4. Discussion
Fig. 3. Photograph demonstrating a loose screw after cyclic testing from a specimen in Group 1.
Treatment of distal humerus fractures in adults remains a challenge in orthopedic trauma surgery, especially in patients with osteopenic bone, articular comminution, and fractures involving the supracondylar region (Jupiter, 2008; O'Driscoll, 2009). Stable fracture fixation is important to allow for early range of motion at the elbow. The current standard of care suggests the use of two plates for optimal stabilization, however, no standard exists with respect to plate type and configuration (Galano et al., 2010; Nauth et al., 2011). The present study attempts to address this controversy by examining two unresolved issues: precontoured locking plates versus nonlocking plates, and perpendicular versus parallel configurations. To our knowledge, this is the only available study that tests these two variables biomechanically using a single protocol.
700
R.C. Koonce et al. / Clinical Biomechanics 27 (2012) 697–701
Table 1 Results of stiffness testing and load to failure. Results are shown for each stiffness test with mean and standard deviation values. There was no statistically significant difference between the three groups in any of the stiffness testing parameters or load to failure.
Group 1 Group 2 Group 3 ANOVA (P)
Anterior bending stiffness (N/mm) Mean (Std. Dev.)
Posterior bending stiffness (N/mm) Mean (Std. Dev.)
Axial stiffness (N/mm) Mean (Std. Dev.)
Torsional stiffness (N * m/Deg.) Mean (Std. Dev.)
Load to failure (N) Mean (Std. Dev.)
1160.1 (221.1) 994.9 (311.1) 1236.9 (283.8) 0.15
923.1 (227.8) 935.9 (164.8) 1051.6 (243.7) 0.35
1430.4 (309.1) 1304.4 (416.8) 1558.0 (386.3) 0.33
2.3 (1.1) 2.0 (0.7) 2.4 (0.8) 0.48
1441.1 (366.9) 1617.6 (280.5) 1548.5 (422.4) 0.55
This human cadaver distal humerus fracture model with metaphyseal comminution was chosen in this study for four primary reasons. First, distal humerus fracture fixation failures and nonunions most commonly occur in the supracondylar region (Jupiter, 2008; O'Driscoll, 2009). Secondly, by analyzing stiffness data and screw loosening in a pure metaphyseal gap model, we have eliminated the confounding variable of intra-articular fracture fixation. Third, with an average cadaver specimen age of 72 years, our study examines fixation in an age group where osteopenia and osteoporosis are common. Last, our testing protocol is based on a previous well-designed and accepted biomechanical fracture model in the distal humerus (Korner et al., 2003, 2004). The authors recognize several limitations in the present study. The ability to visualize distal screw loosening during cyclic loading may have shown a difference in time to screw loosening between groups. Our testing protocol did not allow for this because they were covered by potting material and could not be inspected until after ultimate load testing. Our protocol is also limited to analyzing one fracture pattern. Another limitation is that the study addresses only biomechanical factors, and does not account for the biological factors associated with surgical approach to the distal humerus, plate application, and bone healing. We also did not form an interdigitating arch with our distal screws the parallel plated configuration, which is proposed by other authors to be biomechanically advantageous (O'Driscoll, 2005). Last, our study does not address CRPs in a parallel configuration, nor does it include LCPs. Several prior biomechanical studies have addressed distal humerus fracture fixation with varying protocols, hardware configurations, and results. We identified 10 previous studies that compare plates placed in parallel versus perpendicular configurations. Of these, six studies demonstrated biomechanical superiority of the parallel plate configuration (Arnander et al., 2008; Schemitsch et al., 1994; Self et al., 1995; Stoffel et al., 2008; Windolf et al., 2010; Zalavras et al., 2011), two showed similar biomechanical properties between groups (Kollias et al., 2010; Schwartz et al., 2006), and one demonstrated superiority of perpendicularly placed plates (Jacobson et al., 1997). One
Fig. 4. Distal screw loosening after cyclic loading. Graphical representation of loose screws distal to the osteotomy site after cyclic testing. CRPs (Group 1) had a significantly higher number of loose screws when compared to PDHLPs (Groups 2 and 3).
study showed mixed results with parallel plates stiffer in extension and perpendicular plates stiffer in flexion (Penzkofer et al., 2010). With the advent of angular-stable locking plates, surgeons may assume that these newer and more expensive systems provide more biomechanically stable fracture fixation. We identified four biomechanical studies that examined locking versus non-locking plates. Schuster et al. (2008) compared PDHLPs, LCPs, and CRPs and showed similar stiffness properties between all groups but a lower failure rate in PDHLPs compared to CRPs after cyclic testing. Korner et al. (2004) showed no difference in stiffness between locking plates (using LCPs) and non-locking plates (using CRPs) placed in a perpendicular arrangement. Tejwani et al. (2009) demonstrated that two dorsally applied CRPs were superior to one laterally applied PDHLP. Windolf et al. (2010) showed that a parallel interconnected CRP and a non-locking one-third tubular plate are superior to perpendicular LCPs. Aside from the present study, two other publications compare PDHLPs in both parallel and perpendicular configurations. Stoffel et al. (2008) showed parallel PDHLPs were stiffer in axial compression and external rotation than perpendicular PDHLPs. Penzkofer et al. (2010) showed that parallel PDHLPs had higher stiffness in extension, and that perpendicular PDHLPs were stiffer in flexion. Neither study included a CRP group for comparison. It is clear that the biomechanical data available prior to our study is inconsistent. Our stiffness and load to failure data suggest that in the immediate postoperative period all groups provide similar stiffness and load to failure, which is in agreement with some studies and contradicts others. It is very difficult to compare prior studies to each other and to ours because of wide variations in testing protocols, testing media (cadaver versus sawbones), plate and screw manufacturers, and definitions of failure. Some of the prior studies (Penzkofer et al., 2010; Schuster et al., 2008; Stoffel et al., 2008; Windolf et al., 2010; Zalavras et al., 2011) also evaluated a different fracture pattern (AO/ OTA 13-C2). The inclusion of an intra-articular split provides a confounding variable for evaluating instability in the location most prone to failure—the supracondylar region (Jupiter, 2008; O'Driscoll, 2009). Our study protocol provides useful data for screw loosening after postoperative range of motion is initiated in the first 4000 cycles at 1 Hz. This frequency and number of cycles are comparable to previous studies (Korner et al., 2004; Self et al., 1995), and estimate approximately 3 months of elbow rehabilitation. Distal screw loosening after cycling was significantly lower (P b 0.001) in the PDHLP groups compared to the LCP group. This result was expected given the locking mechanism of all screws in the PDHLP plates. None of the screws proximal to the osteotomy gap failed in any of the three groups. We attribute this to the screws being bicortical and the increased bone mineral density in the diaphyseal portion of the humerus when compared to the distal portion (Diederichs et al., 2009). Schuster et al. (2008) also demonstrated a significantly lower failure rate in screw pullout comparing PDHLPs to CRPs. Based on our screw loosening data and prior studies we draw two conclusions: (a) locking screws are not necessary when placed bicortically in diaphyseal bone; and (b) non-locking screws placed in more osteopenic metaphyseal and
R.C. Koonce et al. / Clinical Biomechanics 27 (2012) 697–701
peri-articular bone have a higher rate of loosening when subjected to cyclic loading. The site-specific PDHLPs have the advantage of a locking mechanism over the CRPS. There is also a theoretical advantage of an additional screw hole in the terminal 33 mm of the plates used in our study (four in the PHLPs versus three in the CRPs). We originally hypothesized that the locking mechanism plus the addition of a fourth screw in each column of osteopenic bone would account for higher stiffness and load to failure in the PDHLP groups. The testing results, however, rejected this hypothesis since there was no statistically significant difference between the three groups in stiffness testing or load to failure. We also hypothesized that specimens with loose screws would have lower loads to failure, but we were unable to demonstrate a difference in specimens with loose screws. We offer two potential reasons for these unexpected results. It is possible that despite screw loosening, the remaining screws in the CRP plate provided enough redundancy to maintain stiffness and yield strength similar to the PDHLP constructs. The non-locking screws in the CRP group also do not require aiming guides and had the advantage of being aimed in the direction that allowed for the longest possible screw, maximizing bone purchase. An alternative hypothesis in favor of the CRP constructs is that they may actually have benefitted from the use of larger diameter screws in the distal fragment (3.5 mm versus 2.7 mm screws).
5. Conclusions Our results demonstrate that perpendicular CRPs have similar stiffness and load to failure properties compared to perpendicular and parallel PDHLPs when tested in a human cadaver distal humerus metaphyseal gap model. However, PDHLPs had a significantly lower rate of screw loosening in the first 4000 cycles of cyclic testing. This study shows that all three tested constructs demonstrate adequate biomechanical fixation for these fractures in the early postoperative period. Lower cost, ease of exposure, and custom shaping of implants are cited as advantages of perpendicular CRPs over parallel PDHLPs (Arnander et al., 2008; Jupiter, 2008; Nauth et al., 2011), and from a biomechanical standpoint our study supports this treatment approach. In order to solve the biomechanical debate regarding optimal fixation, a comprehensive biomechanical study including CRPs, LCPs, and PDHLPs from multiple manufacturers in parallel and perpendicular configurations is needed.
Acknowledgments We thank John Henry Carson for his help in preparation of humerus specimens. We also thank Smith & Nephew, Inc. for providing a research grant and implant hardware used for this project.
701
References Abzug, J.M., Dantuluri, P.K., 2010. Use of orthogonal or parallel plating techniques to treat distal humerus fractures. Hand Clin. 26, 411–421. Arnander, M.W., Reeves, A., MacLoad, I.A., Pinto, T.M., Khaleel, A., 2008. A biomechanical comparison of plate configuration in distal humerus fractures. J. Orthop. Trauma 22, 332–336. Brorson, S., 2009. Management of fractures of the humerus in ancient Egypt, Greece, and Rome. Clin. Orthop. Relat. Res. 467, 1907–1914. Diederichs, G., Issever, A.S., Greiner, S., Linke, B., Korner, J., 2009. Three-dimensional distribution of trabecular bone density and cortical thickness in the distal humerus. J. Shoulder Elbow Surg. 18, 399–407. Galano, G.J., Ahmad, C.S., Levine, W.N., 2010. Current treatment strategies for bicolumnar distal humerus fractures. J. Am. Acad. Orthop. Surg. 18, 20–30. Green, A., 2009. Open reduction and internal fixation with 90–90 plating of bicolumn distal humerus fractures. Instr. Course Lect. 58, 515–519. Jacobson, S.R., Glisson, R.R., Urbaniak, J.R., 1997. Comparison of distal humerus fracture fixation: a biomechanical study. J. South. Orthop. Assoc. 6, 241–249. Jupiter, J.B., 2008. The management of nonunion and malunion of the distal humerus—a 30-year experience. J. Orthop. Trauma 22, 742–750. Kollias, C.M., Darcy, S.P., Reed, J.G., Rosvold, J.M., Shrive, N.G., Hildebrand, K.A., 2010. Distal Humerus internal fixation: a biomechanical comparison of 90° and parallel constructs. Am. J. Orthop. 39, 440–444. Korner, J., Lill, H., Muller, L.P., Rommens, P.M., Schneider, E., Linke, B., 2003. The LCPconcept in the operative treatment of distal humerus fractures—biological, biomechanical and surgical aspects. Injury 34, SB20–SB30. Korner, J., Diederichs, G., Arzdorf, M., Lill, H., Josten, C., Schneider, E., et al., 2004. A biomechanical evaluation of methods of distal humerus fracture fixation using locking compression plates versus conventional reconstruction plates. J. Orthop. Trauma 18, 286–293. Nauth, A., McKee, M.D., Ristevski, B., Hall, J., Schemitsch, 2011. Distal humeral fractures in adults. J. Bone Joint Surg. Am. 93, 686–700. O'Driscoll, S.W., 2005. Optimizing stability in distal humerus fracture fixation. J. Shoulder Elbow Surg. 14, 186–194. O'Driscoll, S.W., 2009. Parallel plate fixation of bicolumn distal humeral fractures. Instr. Course Lect. 25, 521–528. Penzkofer, R., Hungerer, S., Wipf, F., von Oldenburg, G., Augat, P., 2010. Anatomical plate configuration affects mechanical performance in distal humerus fractures. Clin. Biomech. 25, 972–978. Rüedi, T., Buckley, R., Moran, C., 2007. AO Principles of Fracture Management, Second edition. Thieme Verlag, New York. Schemitsch, E.H., Tencer, A.F., Henley, M.B., 1994. Biomechanical evaluation of methods of internal fixation of the distal humerus. J. Orthop. Trauma 8, 468–475. Schuster, I., Korner, J., Arzdorf, M., Schwieger, K., Diederichs, G., Linke, B., 2008. Mechanical comparison in cadaver specimens of three different 90-degree doubleplate osteosynthesis for simulated C2-type distal humerus fractures with varying bone densities. J. Orthop. Trauma 22, 113–120. Schwartz, A., Oka, R., Odell, T., Mahar, A., 2006. Biomechanical comparison of two different periarticular plating systems for stabilization of complex distal humerus fractures. Clin. Biomech. 21, 950–955. Self, J., Viegas, S.F., Buford WL, J.R., Patterson, R.M., 1995. A comparison of double-plate fixation methods for complex distal humerus fractures. J. Shoulder Elbow Surg. 4, 10–16. Stoffel, K., Cunneen, S., Morgan, R., Nicholls, R., Stachowaik, G., 2008. Comparative stability of perpendicular versus parallel double-locking plating systems in osteoporotic comminuted distal humerus fractures. J. Orthop. Res. 26, 778–784. Tejwani, N.C., Murthy, A., Park, J., McLaurin, T.M., Egol, K.A., Kummer, F.J., 2009. Fixation of extra-articular distal humerus fractures using one locking plate versus two reconstruction plates: a laboratory study. J. Trauma 66, 795–799. Windolf, M., Maza, E.R., Gueorguiev, B., Braunstein, V., Schwieger, K., 2010. Treatment of distal humeral fractures using conventional implants. Biomechanical evaluation of a new implant configuration. BMC Musculoskelet. Disord. 11, 172–178. Zalavras, C.G., Vercillo, M.T., Jun, B.J., Otarodifard, K.J., Itamura, J.M., Lee, T.Q., 2011. Biomechanical evaluation of parallel versus orthogonal plate fixation of intra-articular distal humerus fractures. J. Shoulder Elbow Surg. 20, 12–20.
Contents lists available at SciVerse ScienceDirect
Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech
Are conventional reconstruction plates equivalent to precontoured locking plates for distal humerus fracture fixation? A biomechanics cadaver study☆ Ryan C. Koonce a,⁎, Todd H. Baldini b, Steven J. Morgan c a b c
Skagit Regional Clinics, 1400 E. Kincaid Street, Mount Vernon, WA 98274, United States University of Colorado School of Medicine, Department of Orthopaedic Surgery, 13001 E. 17th Place, Aurora, CO 80045, United States Swedish Medical Center, MOTUS Orthopaedics, 701 E Hampden Ave Suite 515, Englewood Co 80113 United States
a r t i c l e
i n f o
Article history: Received 20 November 2011 Accepted 27 March 2012 Keywords: Humerus fracture Fixation Locking plate Biomechanics Cadaver
a b s t r a c t Background: The optimal plate type and configuration for distal humerus fracture fixation has yet to be defined. Available biomechanical studies show conflicting results. No existing studies compare conventional reconstruction plates to newer precontoured distal humerus locking plates in both parallel and perpendicular configurations. Methods: Three groups of humerus specimens were compared via biomechanical testing in a cadaver model simulating metaphyseal comminution. Group 1 consisted of conventional reconstruction plates in a perpendicular configuration. Group 2 used precontoured locking plates in a perpendicular configuration. Group 3 used precontoured locking plates in a parallel configuration. Each group was tested for stiffness in anterior bending, posterior bending, axial compression, and torsion. The specimens then underwent cyclic loading followed by single load to failure in posterior bending. Findings: There was no significant difference between the three groups for anterior bending, posterior bending, axial compression, or torsional stiffness. There was no significant difference in load to failure for any of the three groups. Screw loosening was significantly higher in Group 1 when compared to Groups 2 and 3 after cyclic loading. Interpretation: In the early postoperative period, less expensive perpendicular conventional reconstruction plate constructs provide similar stiffness and load to failure properties to newer precontoured locking plate systems regardless of plate configuration. © 2012 Elsevier Ltd. All rights reserved.
1. Introduction Fractures of the adult distal humerus have been described as grave lesions with poor outcomes since circa 415 BC (Brorson, 2009). Doubleplate osteosynthesis is the current standard for treatment in active adults; however, plate type and configuration are topics of controversy within the literature (Green, 2009; Nauth et al., 2011). Historically, the AO group has recommended treatment with conventional reconstruction plates (CRPs) in a perpendicular configuration (Rüedi et al., 2007); wherein the lateral column plate is placed posteriorly and the medial column plate turned approximately 90° and placed medial to the supracondylar ridge. There is a current trend toward use of precontoured distal humerus locking plates (PDHLPs) in a parallel configuration (Nauth et al., 2011; O'Driscoll, 2005); where plates are placed on the medial and
☆ Location of work: This study was performed at the University of Colorado School of Medicine, Department of Orthopaedic Surgery Biomechanics Laboratory in Aurora, CO, USA. ⁎ Corresponding author. E-mail addresses: [email protected] (R.C. Koonce), [email protected] (T.H. Baldini), [email protected] (S.J. Morgan). 0268-0033/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2012.03.008
lateral columns approximately 180° to each other. These newer plates are attractive as an alternative to CRPs because of angular screw stability and an anatomically precontoured shape, but evidence to recommend the routine use of locking plates over non-locking plates for distal humerus fractures is insufficient (Nauth et al., 2011). Several biomechanical studies with conflicting results have been published comparing CRPs, PDHLPs, and locking compression plates (LCPs) in various configurations. With regard to biomechanical stability of various plate and screw constructs, the following issues are currently unproven: (a) whether newer PDHLPs are superior to CRPs; and (b) whether the parallel plate configuration is superior to the perpendicular configuration. With a lack of consensus on these issues, individual surgeon preference and experience often dictate the choice of implant and implant position for internal fixation of distal humerus fractures (Abzug and Dantuluri, 2010; Schwartz et al., 2006). Currently no single study compares the biomechanical properties of perpendicularly placed CRPs to PDHLPs in both parallel and perpendicular configurations. The aim of our study is to make these comparisons using human cadaver specimens in an established biomechanical testing model (Korner et al., 2003, 2004).
698
R.C. Koonce et al. / Clinical Biomechanics 27 (2012) 697–701
2. Methods 2.1. Specimens Thirty fresh frozen left human upper extremity specimens were obtained (Science Care™, Phoenix, AZ, USA) and stripped of all soft tissue and visually inspected for pathology. The mean age of human cadaver specimens used was 72.3 years (range 51 to 98, standard deviation 10.8). There were 17 humeri from female cadavers and 13 from male cadavers. Total bone mineral density (BMD) of the distal third of each humerus specimen was measured with a dual energy X-ray absorptiometry (DEXA) machine (Hologic, Inc., Bedford, MA, USA). The mean and standard deviation for bone mineral density were mean 0.69 (SD 0.18 g/cm 2). After the DEXA scans were completed the humeri were randomly assigned to three groups of ten. The mean BMDs and standard deviations of the three groups were as follows (g/cm 2): Group 1 (0.69,0.19), Group 2 (0.64,0.22), and Group 3 (0.73,0.12). These were statistically compared using one way ANOVA to ensure there was no significant difference in BMD between the three groups (P = 0.98). The specimens were then wrapped and stored in saline soaked gauze at −20 °C prior to testing. 2.2. Fracture model Specimens were thawed for 12 h at room temperature. First, plates with moldable segments were contoured to fit to the bone. Next, the proximal and distal screw holes of medial and lateral/posterolateral plates were drilled into the bone to ensure an anatomical reduction. A supracondylar distal humerus fracture model was then created using a band saw to cut a transverse 5 mm osteotomy gap just proximal to the olecranon fossa to simulate metaphyseal comminution (OTA/AO type 13-A3.3). The gap was large enough to avoid bone contact between proximal and distal fragments during testing. 2.3. Fracture fixation Three different plating configurations and four plate types were used for fracture fixation. All implants used are Food and Drug Administration approved for distal humerus fracture fixation. Group 1: Non-locking 3.5 mm stainless steel 8-hole (94 mm length) CRPs, (Smith & Nephew, Inc., Memphis, TN, USA), were anatomically mounted around the medial epicondyle on the ulnar column and along the posterolateral surface on the radial column in a perpendicular configuration using 3.5 mm non-locking cortex screws (Fig. 1A, D). Group 2: Plates were placed in a similar manner to plates in Group 1 using 7-hole medial (103 mm) and 7-hole posterolateral (107 mm) stainless steel posterolateral PDHLPs (Smith & Nephew PERI-LOC) using only locking screws in a perpendicular configuration (Fig. 1B, E). Group 3: 7-hole medial (103 mm) and lateral (102 mm) stainless steel PDHLPs (PERI-LOC, Smith & Nephew, Inc., Memphis, TN, USA) were placed on ulnar and radial columns using only locking screws in a parallel configuration (Fig. 1C, F). Plate lengths were chosen to most closely match the overall working distance of the PDHLPs to an 8-hole CRP. Plate thickness is 2.8 mm for all CRPs and 3.1 mm for all PDHLPs. Pilot holes of appropriate size for the screws per the manufacturers' specifications were made with a power drill. Screws were inserted by hand to a tightness of two fingers. A 1.7 Nm torque limiting screw driver was utilized for all locking screw insertions. Proximal to the osteotomy site, three 3.5 mm bicortical screws were placed in every plate. Distal to the osteotomy site, monocortical screws were used to avoid penetration of articular surfaces. CRPs were fixed distally with three monocortical 3.5 mm non-locking screws. PDHLPs were fixed distally with one monocortical 3.5 mm locking screw and three
Fig. 1. Photographs and plain X-rays demonstrating three different testing groups. Group 1 (A and D) = perpendicular CRPs; Group 2 (B and E) = perpendicular PDHLPs; Group 3 (C and F) = parallel PDHLPs.
monocortical 2.7 mm locking screws. Penetration of the olecranon fossa was avoided in all groups. 2.4. Potting the specimens After the humeri were plated they were potted for mechanical testing. The proximal end of the humeri was cut off with a band saw 6 cm from the most proximal end of the plates. Five centimeters of the proximal end of the bone was potted in the center of a 4 in. diameter aluminum tube with Dyna-Cast hard setting urethane (Kindt-Collins Company LLC, Cleveland, OH, USA). Fixing the distal end of the humeri was accomplished by first placing a highly elastic modeling compound (Hasbro, Inc., Pawtucket, RI, USA) over the distal plate and screws. This prevented plates and screws from being rigidly fixed in the potting compound and allowed the screws to back out during testing. The distal end was then potted to within 1 cm of the osteotomy in the center of a 4 in. square tube with Dyna-Cast. 2.5. Mechanical testing Three different types of mechanical testing were performed: stiffness testing, cyclic loading, and single load to failure. All tests were performed on an Instron Model 1321 closed loop servo-hydraulic test machine (Instron Corporation, Canton, MA, USA). All data were collected at 50 Hz on a PC equipped with a Keithley 1802HC (Keithley Instruments, Inc., Cleveland, OH, USA) analog to digital board and TestPoint (Capital Equipment, Corp., Billerica, MA, USA) data acquisition software. Stiffness testing was performed in axial compression (250 N), torsion (+/−1.6 Nm), anterior 4 point bending (4.5 Nm), and posterior 4 point bending (4.5 Nm) (Fig. 2). These load levels were chosen to avoid plastic deformation of the construct while stiffness testing (Korner et al., 2004). Four point bending was used to provide a constant bending moment across the fracture site. All tests were sinusoidal wave forms run for 4 cycles at 0.2 Hz. The stiffness was measured on the 4th cycle between 20% and 80% of the peak applied load.
R.C. Koonce et al. / Clinical Biomechanics 27 (2012) 697–701
699
Fig. 2. Biomechanical testing setup. An Instron Model 1321 closed loop servo-hydraulic test machine (Instron, Corp., Canton, MA, USA) is pictured with a distal humerus specimens following osteosynthesis. The testing setup is shown for (A) axial compression and torsion, and (B) posterior four-point bending.
After the stiffness tests were completed the specimens were cyclically loaded for 4000 cycles at 1 Hz to 4.5 Nm in posterior bending. This frequency and number of cycles is comparable to previous studies, and estimates approximately 3 months of elbow rehabilitation (Korner et al., 2004; Self et al., 1995). All specimens were inspected after cyclic loading for signs of proximal screw loosening, screw or plate fracture, or bone fracture. All specimens were then tested to ultimate load in posterior 4 point bending. Load was applied under displacement control at a rate of 0.1 mm/s until the specimen failed or 5000 N was applied, the limit of the load cell. After ultimate load testing the potting compound on the distal end of each humerus was carefully removed and signs of failure were recorded. Failure was recorded as implant breakage, bone fracture, or plastic deformation. Distal screw loosening was noted by visually inspecting for screw backout (Fig. 3).
2.6. Statistical analysis A prospective power analysis was performed to determine the required sample size (n) with α = 0.05 and power = 0.8 using JMP data analysis software (SAS Institute Inc., Cary, NC, USA). The power analysis determined an n of 10 was sufficient to obtain statistical significance. After testing, a one way analysis of variance (ANOVA) was used to determine if there was a statistical difference (P b 0.05) in recorded values between the groups assuming the data is normally distributed. A Pearson Chi Square test was used to determine if
there were a significantly greater number of loose screws between groups. 3. Results Using one-way ANOVA, there was no significant difference between the three groups for axial (P = 0.33), torsion (P = 0.48), anterior bending (P = 0.15), or posterior bending (P = 0.35) stiffness. After cyclic testing, specimens were loaded to 5000 N in posterior bending. All 30 failed in plastic deformation. No specimens failed by implant breakage or humerus fracture. Load to failure was defined as the elastic limit for each specimen using standard load–displacement curves. There was no statistically significant difference for load to failure between the three groups (P = 0.55). Results for stiffness testing and load to failure are summarized in Table 1. No specimens demonstrated screw loosening in the diaphysis. All 10 of the Group 1 posterolateral CRPs had two distal screws loose and 8 had all three distal screws loose. Fig. 3 shows an example of a loose screw in Group 1. Four of the Group 1 medial CRPs had at least one loose distal screw. Group 2 had one medial plate with a distal screw loose and one posterior–lateral plate with a distal screw loose. Group 3 had only one distal screw loose on a medial plate. A summary of total screw loosening is shown in Fig. 4. Using a Pearson Chi Square test revealed a statistically significant association between screw type (locking vs. non-locking) and screw loosening. The CRP group, using only non-locking screws, was significantly more likely to have loose screws than the PDHLP groups which used only locking screws (P b 0.001). 4. Discussion
Fig. 3. Photograph demonstrating a loose screw after cyclic testing from a specimen in Group 1.
Treatment of distal humerus fractures in adults remains a challenge in orthopedic trauma surgery, especially in patients with osteopenic bone, articular comminution, and fractures involving the supracondylar region (Jupiter, 2008; O'Driscoll, 2009). Stable fracture fixation is important to allow for early range of motion at the elbow. The current standard of care suggests the use of two plates for optimal stabilization, however, no standard exists with respect to plate type and configuration (Galano et al., 2010; Nauth et al., 2011). The present study attempts to address this controversy by examining two unresolved issues: precontoured locking plates versus nonlocking plates, and perpendicular versus parallel configurations. To our knowledge, this is the only available study that tests these two variables biomechanically using a single protocol.
700
R.C. Koonce et al. / Clinical Biomechanics 27 (2012) 697–701
Table 1 Results of stiffness testing and load to failure. Results are shown for each stiffness test with mean and standard deviation values. There was no statistically significant difference between the three groups in any of the stiffness testing parameters or load to failure.
Group 1 Group 2 Group 3 ANOVA (P)
Anterior bending stiffness (N/mm) Mean (Std. Dev.)
Posterior bending stiffness (N/mm) Mean (Std. Dev.)
Axial stiffness (N/mm) Mean (Std. Dev.)
Torsional stiffness (N * m/Deg.) Mean (Std. Dev.)
Load to failure (N) Mean (Std. Dev.)
1160.1 (221.1) 994.9 (311.1) 1236.9 (283.8) 0.15
923.1 (227.8) 935.9 (164.8) 1051.6 (243.7) 0.35
1430.4 (309.1) 1304.4 (416.8) 1558.0 (386.3) 0.33
2.3 (1.1) 2.0 (0.7) 2.4 (0.8) 0.48
1441.1 (366.9) 1617.6 (280.5) 1548.5 (422.4) 0.55
This human cadaver distal humerus fracture model with metaphyseal comminution was chosen in this study for four primary reasons. First, distal humerus fracture fixation failures and nonunions most commonly occur in the supracondylar region (Jupiter, 2008; O'Driscoll, 2009). Secondly, by analyzing stiffness data and screw loosening in a pure metaphyseal gap model, we have eliminated the confounding variable of intra-articular fracture fixation. Third, with an average cadaver specimen age of 72 years, our study examines fixation in an age group where osteopenia and osteoporosis are common. Last, our testing protocol is based on a previous well-designed and accepted biomechanical fracture model in the distal humerus (Korner et al., 2003, 2004). The authors recognize several limitations in the present study. The ability to visualize distal screw loosening during cyclic loading may have shown a difference in time to screw loosening between groups. Our testing protocol did not allow for this because they were covered by potting material and could not be inspected until after ultimate load testing. Our protocol is also limited to analyzing one fracture pattern. Another limitation is that the study addresses only biomechanical factors, and does not account for the biological factors associated with surgical approach to the distal humerus, plate application, and bone healing. We also did not form an interdigitating arch with our distal screws the parallel plated configuration, which is proposed by other authors to be biomechanically advantageous (O'Driscoll, 2005). Last, our study does not address CRPs in a parallel configuration, nor does it include LCPs. Several prior biomechanical studies have addressed distal humerus fracture fixation with varying protocols, hardware configurations, and results. We identified 10 previous studies that compare plates placed in parallel versus perpendicular configurations. Of these, six studies demonstrated biomechanical superiority of the parallel plate configuration (Arnander et al., 2008; Schemitsch et al., 1994; Self et al., 1995; Stoffel et al., 2008; Windolf et al., 2010; Zalavras et al., 2011), two showed similar biomechanical properties between groups (Kollias et al., 2010; Schwartz et al., 2006), and one demonstrated superiority of perpendicularly placed plates (Jacobson et al., 1997). One
Fig. 4. Distal screw loosening after cyclic loading. Graphical representation of loose screws distal to the osteotomy site after cyclic testing. CRPs (Group 1) had a significantly higher number of loose screws when compared to PDHLPs (Groups 2 and 3).
study showed mixed results with parallel plates stiffer in extension and perpendicular plates stiffer in flexion (Penzkofer et al., 2010). With the advent of angular-stable locking plates, surgeons may assume that these newer and more expensive systems provide more biomechanically stable fracture fixation. We identified four biomechanical studies that examined locking versus non-locking plates. Schuster et al. (2008) compared PDHLPs, LCPs, and CRPs and showed similar stiffness properties between all groups but a lower failure rate in PDHLPs compared to CRPs after cyclic testing. Korner et al. (2004) showed no difference in stiffness between locking plates (using LCPs) and non-locking plates (using CRPs) placed in a perpendicular arrangement. Tejwani et al. (2009) demonstrated that two dorsally applied CRPs were superior to one laterally applied PDHLP. Windolf et al. (2010) showed that a parallel interconnected CRP and a non-locking one-third tubular plate are superior to perpendicular LCPs. Aside from the present study, two other publications compare PDHLPs in both parallel and perpendicular configurations. Stoffel et al. (2008) showed parallel PDHLPs were stiffer in axial compression and external rotation than perpendicular PDHLPs. Penzkofer et al. (2010) showed that parallel PDHLPs had higher stiffness in extension, and that perpendicular PDHLPs were stiffer in flexion. Neither study included a CRP group for comparison. It is clear that the biomechanical data available prior to our study is inconsistent. Our stiffness and load to failure data suggest that in the immediate postoperative period all groups provide similar stiffness and load to failure, which is in agreement with some studies and contradicts others. It is very difficult to compare prior studies to each other and to ours because of wide variations in testing protocols, testing media (cadaver versus sawbones), plate and screw manufacturers, and definitions of failure. Some of the prior studies (Penzkofer et al., 2010; Schuster et al., 2008; Stoffel et al., 2008; Windolf et al., 2010; Zalavras et al., 2011) also evaluated a different fracture pattern (AO/ OTA 13-C2). The inclusion of an intra-articular split provides a confounding variable for evaluating instability in the location most prone to failure—the supracondylar region (Jupiter, 2008; O'Driscoll, 2009). Our study protocol provides useful data for screw loosening after postoperative range of motion is initiated in the first 4000 cycles at 1 Hz. This frequency and number of cycles are comparable to previous studies (Korner et al., 2004; Self et al., 1995), and estimate approximately 3 months of elbow rehabilitation. Distal screw loosening after cycling was significantly lower (P b 0.001) in the PDHLP groups compared to the LCP group. This result was expected given the locking mechanism of all screws in the PDHLP plates. None of the screws proximal to the osteotomy gap failed in any of the three groups. We attribute this to the screws being bicortical and the increased bone mineral density in the diaphyseal portion of the humerus when compared to the distal portion (Diederichs et al., 2009). Schuster et al. (2008) also demonstrated a significantly lower failure rate in screw pullout comparing PDHLPs to CRPs. Based on our screw loosening data and prior studies we draw two conclusions: (a) locking screws are not necessary when placed bicortically in diaphyseal bone; and (b) non-locking screws placed in more osteopenic metaphyseal and
R.C. Koonce et al. / Clinical Biomechanics 27 (2012) 697–701
peri-articular bone have a higher rate of loosening when subjected to cyclic loading. The site-specific PDHLPs have the advantage of a locking mechanism over the CRPS. There is also a theoretical advantage of an additional screw hole in the terminal 33 mm of the plates used in our study (four in the PHLPs versus three in the CRPs). We originally hypothesized that the locking mechanism plus the addition of a fourth screw in each column of osteopenic bone would account for higher stiffness and load to failure in the PDHLP groups. The testing results, however, rejected this hypothesis since there was no statistically significant difference between the three groups in stiffness testing or load to failure. We also hypothesized that specimens with loose screws would have lower loads to failure, but we were unable to demonstrate a difference in specimens with loose screws. We offer two potential reasons for these unexpected results. It is possible that despite screw loosening, the remaining screws in the CRP plate provided enough redundancy to maintain stiffness and yield strength similar to the PDHLP constructs. The non-locking screws in the CRP group also do not require aiming guides and had the advantage of being aimed in the direction that allowed for the longest possible screw, maximizing bone purchase. An alternative hypothesis in favor of the CRP constructs is that they may actually have benefitted from the use of larger diameter screws in the distal fragment (3.5 mm versus 2.7 mm screws).
5. Conclusions Our results demonstrate that perpendicular CRPs have similar stiffness and load to failure properties compared to perpendicular and parallel PDHLPs when tested in a human cadaver distal humerus metaphyseal gap model. However, PDHLPs had a significantly lower rate of screw loosening in the first 4000 cycles of cyclic testing. This study shows that all three tested constructs demonstrate adequate biomechanical fixation for these fractures in the early postoperative period. Lower cost, ease of exposure, and custom shaping of implants are cited as advantages of perpendicular CRPs over parallel PDHLPs (Arnander et al., 2008; Jupiter, 2008; Nauth et al., 2011), and from a biomechanical standpoint our study supports this treatment approach. In order to solve the biomechanical debate regarding optimal fixation, a comprehensive biomechanical study including CRPs, LCPs, and PDHLPs from multiple manufacturers in parallel and perpendicular configurations is needed.
Acknowledgments We thank John Henry Carson for his help in preparation of humerus specimens. We also thank Smith & Nephew, Inc. for providing a research grant and implant hardware used for this project.
701
References Abzug, J.M., Dantuluri, P.K., 2010. Use of orthogonal or parallel plating techniques to treat distal humerus fractures. Hand Clin. 26, 411–421. Arnander, M.W., Reeves, A., MacLoad, I.A., Pinto, T.M., Khaleel, A., 2008. A biomechanical comparison of plate configuration in distal humerus fractures. J. Orthop. Trauma 22, 332–336. Brorson, S., 2009. Management of fractures of the humerus in ancient Egypt, Greece, and Rome. Clin. Orthop. Relat. Res. 467, 1907–1914. Diederichs, G., Issever, A.S., Greiner, S., Linke, B., Korner, J., 2009. Three-dimensional distribution of trabecular bone density and cortical thickness in the distal humerus. J. Shoulder Elbow Surg. 18, 399–407. Galano, G.J., Ahmad, C.S., Levine, W.N., 2010. Current treatment strategies for bicolumnar distal humerus fractures. J. Am. Acad. Orthop. Surg. 18, 20–30. Green, A., 2009. Open reduction and internal fixation with 90–90 plating of bicolumn distal humerus fractures. Instr. Course Lect. 58, 515–519. Jacobson, S.R., Glisson, R.R., Urbaniak, J.R., 1997. Comparison of distal humerus fracture fixation: a biomechanical study. J. South. Orthop. Assoc. 6, 241–249. Jupiter, J.B., 2008. The management of nonunion and malunion of the distal humerus—a 30-year experience. J. Orthop. Trauma 22, 742–750. Kollias, C.M., Darcy, S.P., Reed, J.G., Rosvold, J.M., Shrive, N.G., Hildebrand, K.A., 2010. Distal Humerus internal fixation: a biomechanical comparison of 90° and parallel constructs. Am. J. Orthop. 39, 440–444. Korner, J., Lill, H., Muller, L.P., Rommens, P.M., Schneider, E., Linke, B., 2003. The LCPconcept in the operative treatment of distal humerus fractures—biological, biomechanical and surgical aspects. Injury 34, SB20–SB30. Korner, J., Diederichs, G., Arzdorf, M., Lill, H., Josten, C., Schneider, E., et al., 2004. A biomechanical evaluation of methods of distal humerus fracture fixation using locking compression plates versus conventional reconstruction plates. J. Orthop. Trauma 18, 286–293. Nauth, A., McKee, M.D., Ristevski, B., Hall, J., Schemitsch, 2011. Distal humeral fractures in adults. J. Bone Joint Surg. Am. 93, 686–700. O'Driscoll, S.W., 2005. Optimizing stability in distal humerus fracture fixation. J. Shoulder Elbow Surg. 14, 186–194. O'Driscoll, S.W., 2009. Parallel plate fixation of bicolumn distal humeral fractures. Instr. Course Lect. 25, 521–528. Penzkofer, R., Hungerer, S., Wipf, F., von Oldenburg, G., Augat, P., 2010. Anatomical plate configuration affects mechanical performance in distal humerus fractures. Clin. Biomech. 25, 972–978. Rüedi, T., Buckley, R., Moran, C., 2007. AO Principles of Fracture Management, Second edition. Thieme Verlag, New York. Schemitsch, E.H., Tencer, A.F., Henley, M.B., 1994. Biomechanical evaluation of methods of internal fixation of the distal humerus. J. Orthop. Trauma 8, 468–475. Schuster, I., Korner, J., Arzdorf, M., Schwieger, K., Diederichs, G., Linke, B., 2008. Mechanical comparison in cadaver specimens of three different 90-degree doubleplate osteosynthesis for simulated C2-type distal humerus fractures with varying bone densities. J. Orthop. Trauma 22, 113–120. Schwartz, A., Oka, R., Odell, T., Mahar, A., 2006. Biomechanical comparison of two different periarticular plating systems for stabilization of complex distal humerus fractures. Clin. Biomech. 21, 950–955. Self, J., Viegas, S.F., Buford WL, J.R., Patterson, R.M., 1995. A comparison of double-plate fixation methods for complex distal humerus fractures. J. Shoulder Elbow Surg. 4, 10–16. Stoffel, K., Cunneen, S., Morgan, R., Nicholls, R., Stachowaik, G., 2008. Comparative stability of perpendicular versus parallel double-locking plating systems in osteoporotic comminuted distal humerus fractures. J. Orthop. Res. 26, 778–784. Tejwani, N.C., Murthy, A., Park, J., McLaurin, T.M., Egol, K.A., Kummer, F.J., 2009. Fixation of extra-articular distal humerus fractures using one locking plate versus two reconstruction plates: a laboratory study. J. Trauma 66, 795–799. Windolf, M., Maza, E.R., Gueorguiev, B., Braunstein, V., Schwieger, K., 2010. Treatment of distal humeral fractures using conventional implants. Biomechanical evaluation of a new implant configuration. BMC Musculoskelet. Disord. 11, 172–178. Zalavras, C.G., Vercillo, M.T., Jun, B.J., Otarodifard, K.J., Itamura, J.M., Lee, T.Q., 2011. Biomechanical evaluation of parallel versus orthogonal plate fixation of intra-articular distal humerus fractures. J. Shoulder Elbow Surg. 20, 12–20.