Authors: Mohammad Khak, Jeffrey J. Olson, Patrick Williamson, Mohammad Javad Shariyate, Ahmad Hedayatzadeh Razavi, Kaveh Momenzadeh, Mohammadreza Abbasian, Nadim Kheir, Edward K. Rodriguez, Ara Nazarian
Categories: Research Article, Proximal humerus fracture, Unicortical, Subchondral, Screw, Biomechanics, Osteoporosis, Fracture stability
Source: Heliyon
Authors: Mohammad Khak, Jeffrey J. Olson, Patrick Williamson, Mohammad Javad Shariyate, Ahmad Hedayatzadeh Razavi, Kaveh Momenzadeh, Mohammadreza Abbasian, Nadim Kheir, Edward K. Rodriguez, Ara Nazarian
Proximal humerus fractures are common in adults aged 65 and older and provide challenges for osteoporotic patients due to the risk of suboptimal fixation and complications. Locking plates are often utilized to treat two-part fractures; however, ongoing concerns about their stability exist. This pilot study investigates the biomechanical impact of subchondral locking screws compared to unicortical screws in osteoporotic two-part cadaveric proximal humerus fractures.
Using dual-energy X-ray absorptiometry (DXA), cadaveric shoulder specimens from eight female Caucasian donors with comparable bone mineral densities were used for the study. Either unicortical or bicortical locking screws (the latter representing subchondral screws in real surgeries) were utilized to fix locking plates. Axial load to failure and cyclic physiologic abduction moments were applied in biomechanical testing.
The study found no statistically significant difference in interfragmentary displacement between the unicortical and bicortical groups (p = 0.78). The mechanical properties of both groups were found to be comparable in terms of yield (p = 0.59), ultimate (p = 0.86), and fracture strengths (p = 0.70). Furthermore, rigidity analysis did not identify any significant difference between the two groups (p = 0.22).
Our findings indicate that there is little to no difference in the stability of the construct for osteoporotic two-part proximal humerus fractures, in contrast to general recommendations against unicortical screws. This pilot study suggests that the choice between unicortical and subchondral locking screws may not significantly affect biomechanical characteristics in osteoporotic two-part proximal humerus fractures, despite the study's limitations.
Proximal humerus fractures rank as the third most prevalent fracture among those aged 65 years and above [1], constituting 10 % of all fractures in this population [2]. This injury occurs with high frequency in patients who have been diagnosed with osteoporosis, with implications concerning the purchase of screws with internal fixation.
Osteopenia of the humeral head may result in suboptimal fixation, impaired fracture healing, and screw cutout, leading to revision surgery [3]. Similarly, osteoporosis poses challenges with internal fixation of proximal humerus fractures, particularly concerning implant cutout and varus failure [[4], [5], [6]]. Kristiansen et al. demonstrated that over 50 % of elderly individuals treated with nonlocking plates have fair or poor outcomes owing to screw loosening and pulling out from the humeral head [7]. Thus, locking plates have become the gold standard treatment of proximal humerus fractures [8,9].
Fortunately, even in elderly, osteoporotic patients, reasonable functional outcomes of two-part fractures can be obtained with open reduction and internal fixation (ORIF) and preservation of the humeral head [10] to provide the patients with a stable shoulder with an early range of motion and reduced pain [11].
Although previous studies have shown that the use of bicortical locking screws significantly improves implant stability [12], screw penetration of the subchondral bone in the proximal humerus is not without consequence. Screws that extend beyond the subchondral bone may adversely affect the articular cartilage resulting in accelerated degeneration, either by direct articular damage via chondrolysis or indirect damage via changes in the biomechanical and structural properties of the subchondral bone [13].
Therefore, in this pilot study, we investigated the stabilization of two-part proximal humerus neck fractures using a locking plate placed laterally and secured with either unicortical or bicortical locking screws (representing subchondral screws in actual surgeries) in osteoporotic cadaveric specimens. The unicortical locking screws purchased the trabecular bone below the subchondral bone, and the bicortical locking screws purchased the subchondral bone with minimal penetration into the articular surface. The humeral heads were then subjected to cyclic physiologic abduction loading and axial load to failure.
We hypothesized that bicortical locking screws purchased within the subchondral bone would offer 1) improved fracture stability during physiologically relevant cyclic loading and 2) higher load to failure in comparison to transfixing screws that only engage cancellous bone.
Cadaveric shoulder specimens from 8 female Caucasian donors 74.3 ± 8.4 years of age with BMI values of 21.9 ± 6.2 were obtained (Medcure, Providence, RI, USA) [Table 1]. All specimens underwent dual-energy X-ray absorptiometry (DXA) to verify similar areal bone mineral density (vBMD, mgHA/cm^3^, Fig. 1) for equal distribution between two groups (unicortical fixation and bicortical fixation). Cadaveric studies do not require any approval from the Internal Review Board at Beth Israel Deaconess Medical Center.Table 1Basic Demographics of cadavers.Table 1Age (mean, SD)74.38.4Age (mean, SD)21.96.2Sex (N, % Female)8100Race (N, %) Caucasian8100Fig. 1vBMD values for the unicortical and bicortical fixation groups.Fig. 1
A standard deltopectoral approach was used to expose the proximal humerus. The subdeltoid recess was bluntly developed. A 3.5 mm Synthes proximal humerus locking plate was used with five 3.5 mm locking screws placed into the humeral head and three 3.5 mm non-locking cortical screws into the humeral shaft. For the unicortical group, a 1.25 mm K-wire was inserted through a wire guide to abut the subchondral bone and was then measured. Screws 2 mm shorter were then placed, achieving unicortical purchase only. Based on the manufacturer's guidelines, we applied a 3 Nm torque for the locking screws. To achieve bicortical fixation for the bicortical group, a screw that was 4 mm longer was inserted in a way that it only purchased the subchondral bone with minimal penetration into the articular surface. By utilizing multiple views with the assistance of a C-arm and exposing the articular surface, given the cadaveric nature of the study, we ensured there was no articular penetration. (Fig. 2).Fig. 2Fluoroscopic images of proximal humerus fixation of two-part surgical neck fractures demonstrating a) unicortical fixation and b) bicortical fixation with minimal screw tip penetration.Fig. 2
An osteotomy using a sagittal saw was then performed at the surgical neck to create a 10 mm interfragmentary gap to simulate a comminuted unstable fracture. The scapula and humerus were then potted in two-part epoxy (Smooth-Cast 300q and PMG Smooth-on Inc., Macungie, PA, USA) and loaded onto the biomechanical testing device.
Specimens were tested using a previously validated 7-degree of freedom (DoF) cadaveric shoulder testing system similar to the model simulating physiologic loading of the proximal humerus described by Zettl et al. [14] (Fig. 3). Anatomically positioned cadaveric rotator cuff muscles were affixed around the glenohumeral joint, each tensioned to 25N to stabilize and balance the joint, replicating normal physiologic conditions. The humerus was mounted onto the robotic arm of the 7DOF testing system, allowing for passive glenohumeral abduction in the scapular plane. Cyclic loading, targeting 200 cycles from 45° to 90° of glenohumeral abduction per condition, was performed passively. Rigid body motion (position and orientation) between the two fragments was recorded at 240 Hz using electromagnetic sensors (Liberty, Polhemus, Colchester, VT, USA - Accuracy: 0.76 mm and 0.15°) mounted and rigidly affixed to the proximal humerus and humeral shaft. The change in position between the sensors was recorded as the interfragmentary displacement. A 6DoF load cell (Mini58 IP65/IP68, ATI, Apex, NC, USA) measured joint reactive force during motion beneath the scapula. Interfragmentary displacement (mm) was chosen as the primary outcome measure as it directly correlates with interfragmentary strain, an important biomechanical determinant of fracture healing.Fig. 3a) Presentation of the system simulating the physiologic loading of the glenohumeral joint in abduction with electromagnetic markers to detect interfragmentary gap motion, and b) the axial load-to-failure testing system.Fig. 3
After completing the cyclic loading regimen, the specimens were subjected to a single load-to-failure test. Specimens were mounted onto an Instron 8511 load frame (Instron, Inc, Norwood, MA, USA) and underwent axial loading at 1.5 Hz to failure. The force-displacement graphs from the test were used to calculate and compare the mechanical properties, including yield, ultimate, and strengths (MPa), displacements (mm), and stiffness (N/mm). The final failure mode was assessed during the load-to-failure testing. Failure modes were evaluated through video recordings of each specimen post-testing, and the moment of failure was determined from the force-displacement curve generated during testing.
This study was conducted as a pilot to determine the sample size for a definitive larger study and to assess the mechanical testing method. Statistical analysis was performed using GraphPad Prism (version 9.3.1 for Windows, GraphPad Software, San Diego, CA, USA). The Shapiro-Wilk test was utilized to assess the data distribution. Depending on the data distribution, either an unpaired samples t-test or a Mann-Whitney test was used to compare mechanical properties between the two conditions. Two-tailed p-values less than 0.05 were considered significant.
Maximum Interfragmentary displacement was compared between the two groups, where the unicortical group showed an average displacement of 1.79 ± 1.04 mm, and the bicortical group had a slightly higher average displacement of 1.97 ± 0.57 mm with less variation. However, the difference wasn't statistically significant (P = 0.78), indicating that both groups performed similarly regarding movement at the fracture site (Fig. 4). The mechanical properties, including yield, ultimate, and fracture strengths and displacements for each sample under axial load to failure test, are presented in Fig. 5, with no statistical differences observed between the unicortical and bicortical groups (P > 0.05, 612 MPa ± 409, 1030 MPa ± 898 and 886 MPa ± 808 | 803 MPa ± 550, 1135 MPa ± 804 and 1116 MPa ± 808, respectively). Average stiffness (Fig. 6) was 188 MPa ± 90 for the unicortical group and 111 MPa ± 70 for the bicortical group, with no significant difference (P = 0.22). The observed failure modes were primarily screw pull-out and bone fracture. Notably, no cases of implant failure were observed.Fig. 4Interfragmentary gap displacement between the unicortical and cortical fixation groups.Fig. 4Fig. 5The mechanical properties, including yield, ultimate, and fracture strengths and displacements for both groups.Fig. 5Fig. 6Stiffness values between the unicortical and cortical fixation groups.Fig. 6
The following failure mechanisms were observed during the load-to-failure 1. Screw Pull-Out: This occurred when the screws lost purchase within the bone, resulting in disengagement from the surrounding bone tissue. 2. Bone Fracture: Bone fracture was observed in cases where the applied load exceeded the structural integrity of the osteoporotic bone (Table 2). The fracture typically occurred around the screw insertion points or along the humeral shaft, reflecting the bone's inability to distribute the stress effectively under load. Importantly, no implant-related failures (e.g., plate bending, screw breakage) were noted.Table 2Detailed mechanical results comparing the two groups (unicortical vs. bicortical screws) in terms of fixation strength and stability.Table 2StrengthUnicorticalBiocorticalDisplacement (mm)MeanMinMaxSTDMeanMinMaxSTDP-value**Yield Strength (GPa)0.610.191.080.410.800.311.590.550.59Yield Displacement (mm)4.551.608.313.078.015.469.981.880.10Ultimate Strength (Gpa)1.030.242.230.891.140.441.790.800.86Ultimate Displacement (mm)6.032.339.383.409.817.2410.951.730.11Fracture Strength (Gpa)0.880.241.980.811.120.412.220.810.71Fracture Displacement (mm)**6.582.4610.383.6210.077.5411.271.7010.13
Despite the current standard of care advocating for the utilization of locking plates, the literature reveals a failure rate ranging from 8 % to 22 % [[15], [16], [17]]. Thus, it is possible to identify a potential biomechanical cause of failure. Previous studies have found that double plate fixation [18], medial support screws [19], medial wall reconstruction [20], and the use of locking screws rather than cortical screws [21] enhance stiffness and reduce mechanical failure. In this study, we evaluated the stabilization of two-part proximal humerus neck fractures using a locking plate placed laterally and secured with either unicortical or bicortical locking screws (representing subchondral screws in actual surgeries) in osteoporotic cadaveric specimens. Following reduction and fixation, the humeral heads were subjected to cyclic physiologic abduction loading and axial load to failure.
The manufacturer's guidelines for locking plates recommend using only unicortical screws; however, the literature does not support this assertion. Multiple investigations on the biomechanics of locking screws have determined that the stability of standard unicortical screw fixation is generally inferior to that of bicortical screw fixation; this difference becomes even greater when torsional stresses are considered [[22], [23], [24]]. Previous biomechanical studies primarily examined the mechanical characteristics of bicortical and unicortical locking screws in other fractures, such as femoral neck fractures [25], proximal tibia fractures [26], and fractures of the radius shaft and distal radius [27,28]. Douglass et al. observed that bicortical screw placement might offer a biomechanically superior construct for the stabilization of unstable proximal tibia fractures compared to unicortical screw placement [26]. However, for comminuted radius shaft fractures, Overturf et al. demonstrated no statistically significant differences between the mechanical properties of bicortical and unicortical abutting locking screw fixation techniques [27]. Neder Filho et al. observed comparable outcomes in cases of distal radius fracture [28].
In this study, we used osteoporotic cadaveric specimens to replicate current clinical practice in our investigation. Studies have shown that in nonrigid proximal humerus fracture-fixation techniques, fixation strength was not significantly associated with bone mineral density [29]. Nonetheless, studies using interlocking intramedullary (IM) nails or locking and nonlocking plates as a rigid fixation demonstrate a strong correlation between fixation stability and BMD [30,31]. Consistent with findings from previous studies, screw pull-out was most frequently observed in cases where bone quality was compromised due to osteoporosis, leading to a reduced capacity to withstand mechanical forces. Similarly, in our study, the predominant failure modes identified were screw pull-out and bone fracture.
Though surprising, the ultimate and fracture strength are not different between the bicortical and unicortical fixation groups. In materials with low ductility, such as bone, the drop after the ultimate strength peak is minimal, leading to fracture occurring almost simultaneously. Regarding literature, this outcome is attributed to the osteoporotic nature of the specimens, as the bone was the weakest component in the construct and lacked the mechanical strength necessary to withstand the applied loads.
The results revealed minimal to no significant differences in strength parameters (yield, ultimate, and fracture) and displacement between the unicortical and bicortical groups. These findings further emphasize that the primary failure mechanisms were governed by the weaker osteoporotic bone, which served as the limiting factor in the mechanical performance of the construct. These findings suggests that the depth of screw penetration, whether unicortical or bicortical, does not significantly influence these biomechanical parameters in the tested cadaveric shoulder specimens. Furthermore, the stiffness comparison, an essential factor in understanding the stability and resilience of the fixation, also showed a similar trend, with no differences observed between the two groups. This equivalence in mechanical strength and stiffness, regardless of screw length and penetration depth, challenges the conventional preference for subchondral fixation in the context of proximal humerus fractures. It underscores the potential efficacy of unicortical fixation in clinical settings, offering a viable alternative with comparable biomechanical properties. These results suggest that placing a subchondral locking screw in the fracture fixation of the proximal humerus may not provide a mechanical advantage. However, it could increase the risk of articular penetration, potentially leading to revision surgery and associated complications. This study aimed to evaluate the power necessary for future studies that involve more clinically relevant fracture types. Given the limitations of the current sample size, it is crucial to conduct studies with larger and more diverse sample populations. Such studies would provide more robust and reliable data, allowing for more accurate conclusions regarding the effectiveness of different fixation methods. Larger sample sizes would also enhance the statistical power, making it possible to detect subtle differences and draw more definitive conclusions applicable to clinical practice.
The small sample size and the limitations intrinsic to any mechanical fracture model utilizing cadaveric specimens constitute the limitations of our study. Moreover, the screw depth variation (working length) between the groups could potentially affect the stability results. The study only assessed the fixation of a specific type of simple proximal humeral fracture (most closely mimicking a surgically treated comminuted two-part proximal humerus fracture), and the "fracture" itself was an idealized representation in the form of an osteotomy. This study did not include complex or highly comminuted fractures; hence, these findings should not be extended to more severe fracture patterns. The axial loading of the fractures in the model may not accurately reflect the stresses acting on the proximal humerus. Still, the loading regimen is consistent with the approaches used in other studies [32,33]. Furthermore, our model does not consider the impacts of soft tissues, including the joint capsule, ligaments, and rotator cuff.
Our findings indicate that there is little to no difference in the stability of the construct for osteoporotic two-part proximal humerus fractures, in contrast to general recommendations against unicortical screws. This pilot study suggests that the choice between unicortical and subchondral locking screws may not significantly affect biomechanical characteristics in osteoporotic two-part proximal humerus fractures, despite the study's limitations.
Mohammad Khak: Writing – original draft, Methodology, Formal analysis, Data curation. Jeffrey J. Olson: Writing – original draft, Methodology, Funding acquisition, Data curation, Conceptualization. Patrick Williamson: Writing – review & editing, Methodology, Formal analysis, Data curation. Mohammad Javad Shariyate: Writing – review & editing, Methodology, Funding acquisition, Formal analysis. Ahmad Hedayatzadeh Razavi: Writing – review & editing, Methodology, Formal analysis. Kaveh Momenzadeh: Writing – review & editing, Methodology, Funding acquisition, Formal analysis. Mohammadreza Abbasian: Writing – review & editing, Methodology, Formal analysis, Data curation. Nadim Kheir: Writing – review & editing, Methodology, Funding acquisition, Formal analysis. Edward K. Rodriguez: Writing – review & editing, Methodology, Funding acquisition, Conceptualization. Ara Nazarian: Writing – review & editing, Supervision, Resources, Methodology, Funding acquisition, Conceptualization.
The authors did not use generative AI tools in preparing this work.
•Orthopaedic Research and Education Foundation, Resident Clinician Scientist Training Grant (JJO)•The Joe Fallon Research Fund and the Dr. Louis Meeks BIDMC Sports Medicine Trainee Research Fund at BIDMC Orthopaedic Surgery Department (AN)
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.