Bicortical Compression and Construct Stability with Variable Pitch Locking Screws in Cadaveric Specimens

Abstract

OBJECTIVES:

A variable pitch locking screw is intended to provide interfragmentary compression combined with fixed angle stability of locking plate constructs. The objective of this study was to compare variable pitch locking screws (3.5 mm KreuLock Ti Locking Compression Screws, Arthrex Inc., Naples FL) with standard locking screws (from same manufacturer) in bicortical fixation scenarios in cadaver bone, by assessing: (1) interfragmentary compression and plate-bone compression; and (2) construct biomechanical stability.

METHODS:

Nine matched pairs of fresh-frozen cadaveric specimens with an average age of 67.2 years (range, 37-83) were used. Interfragmentary compression and plate-bone compression associated with insertion of single bicortical screws were compared between the variable pitch and standard locking screws at increasing levels of torque. The specimens tested were distal tibiae having a simulated longitudinal fracture. Fibulae were osteotomized to create a stable longitudinal fracture pattern and were fixed with a five-screw plate construct with either all variable pitch or all standard locking screws. One of the five screws was placed across the osteotomy without lagging. Fibulae were tested cyclically with axial with torsional loading to compare, displacements, rotation and loads at failure, or tested in four-point bending to compare construct stiffness and maximum force to failure.

RESULTS:

Interfragmentary and plate-bone compression forces in the distal tibia model varied across specimens but were significantly higher with variable pitch locking screws compared to standard locking screws [512 N (SD = 324 N) vs. 79 N (SD = 64 N), p = 0.002 & 242 N (SD = 119 N) vs. 104 N (SD = 123 N), p = 0.028, respectively]. In cyclic loading of fibula constructs, no significant differences were detected in construct axial displacement or angular displacement (p > 0.05). In four-point bending, no differences were detected in maximum force or bending stiffness (p > 0.05).

CONCLUSIONS:

Variable pitch locking screws produced interfragmentary compression between cortices and plate-bone compression that were greater than standard locking screws. In a stable bicortical fibula fixation scenario under external loading, the stability of variable pitch locking screw constructs was similar to constructs with standard locking screws.

INTRODUCTION:

Incidences of osteoporotic and neuropathic fractures are increasing in the aging population ^1-6. These cases, involving impaired bone microarchitecture and density, can present challenging fixation and show increased complication rates ^2,4,7-11. Recent reported rates of inadequate fixation for lower leg and ankle fractures have reached 28%, with most re-operations unrelated to infection ^8,12-16. Surgically related drivers of fixation failure include malreduction, inappropriate construct stiffness, and poor screw purchase ^10,17. These issues not only affect fracture healing, but are associated with patient-reported pain, joint stiffness, physical debilitation, and revision surgery ^10,18.

Locked plates are often used for fixation of fractures with compromised bone quality. Locking screws angularly stabilize the fracture through the threaded rigid connection between the plate and screw, and are not intended to compress or reduce fragments ^19-21. In some scenarios, interfragmentary compression combined with locked plating is desirable. Maintaining stability through fragment compression in locking plate constructs can provide an improved environment for primary healing ^22-24. This may be achieved using a separate lag screw combined with a locked plate. Using a locking screw for lagging-by-technique through the plate’s holes also blends these two fixation approaches, but requires more surgical time and expertise²⁵. Furthermore, in scenarios involving complex geometry and poor bone quality, over-drilling may compromise the bone ^22,26.

A novel approach for combining compression with locked plating is variable pitch locking screws (VPLS) that possess both threaded locking at the screw head, and variable pitch threading along the shaft (Fig. 1A). Variable pitch screws, characterized by the pitch increasing moving further from the screw head, can create a compressive effect between fragments along the screw shaft, serving a similar function as lagging techniques ^22,26-34. During insertion and tightening of a variable pitch screw across a fracture, the two bone fragments become engaged with different portions of the screw having different thread pitches. With the distal portion of the screw (furthest from the screw head/drive point) having a larger pitch than the proximal portion, clockwise rotation of the screw results in a translation of the distal fragment toward a fixed proximal fragment. Variable pitch headless compression screws are often used in fixation of scaphoid fractures, and variable pitch screws have been studied in unicortical or cancellous applications ^{22,26,28,32,34-36}. They have also been reported to produce a more gradual increase to peak compression and a less distinct loss of compression compared to lag-screws ²². Although a number of studies have demonstrated their compressive capabilities ^27,35-37, variable pitch threading has been reported to reduce pull-out strength in a synthetic bone model. This may be due to partial bone thread damage during screw insertion/advancement as the different pitches of the screw sequentially engage with the same regions of bone, presumably creating different or larger thread cuts in the bone ³².

Figure 1:

(A) The three screws used in the study: 20 mm length VPLS used for fibula fixation, and the 50 mm VPLS and SLS used for tibia fixation. Tick marks on the scale represent 1 mm increments. Setups for tests of compression involving (B) the placement of the plate and load cell, (C) clamp reduction of 150 N prior to screw placement, and (D) the torque wrench being used to apply the screw. Also shown are specimens after clamp removal in which (E) interfragmentary compression, and (F) plate-bone compression were measured. (G) The different holes used for testing with the clamp covering the unused oblong hole.

Adding a locking feature to variable pitch screws, VPLS can lock with a plate and obviate the need for additional surgical drilling required for lagging-by-technique. The first objective of this study was to quantify the interfragmentary and plate-bone compression in a simulated longitudinal distal tibia fracture using VPLS, and to compare this with standard locking screws (SLS). The SLS, which theoretically should not provide any additional interfragmentary compression after initial reduction, served as an experimental control. The null hypothesis for this objective was that there is no difference in interfragmentary and plate-bone compression between VPLS and SLS in a simulated fracture. The second objective was to compare the construct stability of specimens fixed with a plated construct with either all VPLS or all SLS. For these tests, longitudinal fractures were created in fibula specimens, and five-screw constructs were created with one screw spanning the fracture. The null hypothesis for this objective was that there is no difference in construct stability between VPLS and SLS in plated constructs.

METHODS:

Specimens:

Nine matched pairs of fresh-frozen cadaveric below-knee specimens without previous deformities were obtained (Anatomy Gifts Registry, Hanover, MD). The cohort included 5 male and 4 female donors with an average age of 67.2 years (range, 37-83) and body-mass-index of 28.6 (range, 18-38). All specimens were dissected and stripped of all soft-tissue, sprayed with saline, and sealed in airtight bags between preparation steps, and underwent no more than two freeze-thaw cycles prior to testing. Compression testing was conducted with distal tibiae, whereas construct stability testing was conducted with proximal and distal fibulae.

Compression Testing:

Interfragmentary compression tests were performed on 16 limbs from 9 unique donors, and plate-bone compression tests were performed on 14 limbs from 9 unique donors. Tibiae were cut in the coronal plane through the middle of the shaft using a bandsaw to separate the anterior and posterior halves. The anterior half of each tibia was fastened to a fully constrained aluminum fixture using a U-clamp. After removing the proximal-posterior segment of bone, the distal-posterior segment was returned to its anatomical position and a 7-hole titanium straight locking plate (AR-9943C-07, Arthex Inc., Naples FL) was laid on the flat posterior aspect of the distal metaphyseal tibia (Fig. 1B). The selected plate possessed a compression/positioning non-locking oblong hole (Fig. 1). Using the recommended drill guide, four 2.5 mm holes were drilled, one at each end hole of the plate and in the two central locking holes. The adjacent middle holes were used for separate tests, one for each screw type, to limit variability in cortical thickness and bone density between tests. The plate was loosely positioned with 2 mm k-wires to maintain alignment. A calibrated 4 kN capacity (± 2 N accuracy, 0.01 N resolution), liquid resistant load-cell with a central thru-hole (7.1 mm thickness, FSH03959, FUTEK Inc., Irvine, CA) was then placed between distal-posterior and distal-anterior segments and two loading washers (1.5 mm thickness), with the central alignment k-wire passing through (Fig. 1) ^22,36,37.

Using clamps on either side of the central alignment k-wire, the plate was clamped onto the bones with a compressive force of 150 N. This clamping force was determined through pilot tests overseen by the orthopaedic surgeon on the project. Because of the presence of a fracture between cortices, a clamping force was needed when the screw tip (of either screw type) entered into the far cortex, to prevent the screw from pushing the far cortex away from near cortex. Although such clamping forces are typically not known clinically, controlling it ensured equal initial compression amongst specimens ^28,38,39 (Fig. 1C). A single author, blinded to screw-type, was then handed a randomly assigned 50 mm long screw which was either a (1) VPLS (3.5 mm KreuLock Ti Locking Compression Screw, Arthrex Inc., Naples, FL), or (2) a SLS (3.5 mm Low Profile Ti Locking Screw, Arthrex Inc., Naples, FL). Both screws possessed the same locking head feature with a 1 mm pitch and dual helix thread path. The pitch of the SLS screw and initial pitch (closest to head) of the VPLS were also both 1 mm (Supplemental Materials). The screw was forwarded downward through the guide-holes using the recommended driver until the locking head threads initially contacted the plate. Then, a 12 N∙m capacity digital torque wrench (Westward Mini Digital Torque Wrench, 6PAE8) with an accuracy of ± 3% and resolution of 0.01 N∙m was used to advance the screws to specific torque levels. Load cell measurements were obtained (after removing the torque wrench) at the following stages: (1) initial screw-head contact with the plate, (2) 1 N∙m, (3) 2 N∙m, (4) 3 N∙m of torque and (5) clamp release ⁴⁰. In pilot testing, 3 N∙m of torque was determined to be sufficient in providing a flush surface between the screw and the top of the plate (Fig. 1D-E).

After testing interfragmentary compression associated with the first screw, the screw was removed, and the other screw type (VPLS or SLS) was tested in the adjacent locking hole in the same manner as above. It was assumed that the testing of one screw did not affect the testing of the other screw in a different hole, and use of the same bone provided a more statistically powerful paired comparison. After testing this second screw and removing it, plate-bone compression tests were conducted; the plate was translated distally, and both screw types were tested with the load cell positioned instead between the plate and the bone (Fig. 1F). Each bone was thus tested 4 times producing 4 adjacent holes 25 mm apart, including two interfragmentary compression tests and two plate-bone compression tests.

Stability Testing - Specimen Preparation:

Fibulas were first cut in the midshaft in the transverse plane, separating the proximal and distal segments. Distal specimens were used for cyclic combined loading, and proximal specimens were used for four-point bending⁴¹. For both fibular halves, a micro-oscillating bone saw was used to make z-osteotomies (40 mm length) that would allow for plate placement on the flattest portion of the posterior segment, parallel to the coronal face of the cut (Fig. 2A) ⁴². After cuts were made, fibular specimens were potted in polyvinyl chloride (PVC) pipe using a urethane filler (Masters DYNA CAST; Freeman Manufacturing and Supply Co) and two additional high-hold steel screws were threaded radially through the potted PVC for additional fixture rigidity ^43,44. Then the z-osteotomies were anatomically reduced with a 7-hole Titanium straight locking plate placed over the flat posterior side. A specimen was fixed with a construct that included either all VPLS or all SLS. The contralateral side received the other screw type, and the author applying the fixation was blinded to the assignments. A plate clamp was applied on either side of the central hole, overlying the osteotomy, and five holes were drilled using the appropriate drill bit and guide. A single screw, either a VPLS or SLS, spanned the osteotomy while the other two pairs of the same screw type were located proximal and distal to the osteotomy. The screws were torqued to 3 N∙m in the same sequential order for all specimens; all screws were partially threaded into the near cortices first, followed by full tightening of the middle screw, and then tightening of the outer screws.

Figure 2:

Biomechanical setups for the external loading of fibulas. (A) The cyclic loading specimen and (B) setup. (C-D) The four-point bending setup and failure of a specimen under loading.

Stability Testing - Cyclic Loading:

Cyclic loading tests were performed on 7 matched pairs of distal fibulas. The potted specimens were fastened to the vertical actuator and baseplate of a servo-hydraulic test frame (Bionix 370.02; MTS, Eden Prairie, MN) using set screws to align the fibula axis with the vertical axis of the actuator (Fig. 2B). The distal fixture was fully constrained while the actuator applied a simultaneous torsional and axial load to the upper fixture with all other degrees of freedom fixed. The compressive load and torque were measured simultaneously with a calibrated load cell (662.20H-05 with 25 kN and 250 N∙m capacities, 0.22% and 0.021% accuracies, and 0.003 N and 0.00003 N∙m resolutions respectively; MTS, Eden Prairie, MN) located below the baseplate.

With zero force and torque on the specimen, the protocol began with a 20-second ramp to a pre-load of 20 N and 1 N∙m. Cyclic loading was applied in a stepwise fashion every 100 cycles at 1 Hz with the pre-load values representing the valleys of the sinusoidal loading. The peak magnitudes started with 40 N and 1.5 N∙m of force and torque respectively for the first 100 cycles (first loading stage). These peaks were increased by 20 N (force) and 0.5 N∙m (torque) for each subsequent stage. Specimens were tested until failure which was indicated by inability to resist peak force or torque, or the end of the final stage with peaks of 180 N force and 5 Nm torque. At each cycle the peak construct axial displacement and rotational angle were acquired by the MTS system. Values only prior to stage of failure were used for analyses.

Stability Testing - Four-Point Bending:

Four-point bending tests were performed on 6 matched pairs of proximal fibulas. The potted proximal fibula constructs were placed into a four-point bending fixture with the posteriorly placed plate facing upwards (Fig. 2C). The setup parameters were based on ASTM F382 standard for orthopaedic plates , however, in this case we tested a construct with bone⁴⁵. Loads were applied through high-strength cylindrical steel roller bearings. The lower support rollers and the upper loading rollers had spans of 195-mm and 145-mm respectively. The vertical load was applied through a rotational bearing under displacement control at a rate of 0.01 mm/sec until failure or prolonged plastic deformation of the plate was observed visually and, in the force-displacement curve (Fig. 2D). The maximum force was taken as the final point before a 5% drop in force, and the stiffness was determined as the slope of a linear fit of the plotted force vs. displacement points acquired at a sampling rate of 30 Hz.

Bone Mineral Density:

After compression tests were complete, thawed tibiae underwent computed tomography scans in a clinical scanner (0.7 mm slice spacing)⁴⁶. A phantom with five different known hydroxyapatite densities was scanned with the bones for image calibration. Images were analyzed in Avizo software (Thermo Fisher Scientific, Waltham, MA). A volume region that captured 20 mm of bone length between two adjacent screw holes was analyzed. The mean density (mg HA/cm³) of this region was determined, including cortical and trabecular bone, and any empty intramedullary space.

Statistical Analysis:

For analysis of compression test results, the outcomes of interest were the interfragmentary and plate-bone compression forces at the final unclamped stage. For each outcome, a linear mixed-effects model, which accounts for correlated data, was used to compare the standard vs. variable pitch locking screw types. Data from the two screw types in the same limb in adjacent holes were treated as paired data. Data from both limbs of the same donor were treated as correlated measurements. All statistical significance was set at α = 0.05. For analysis of cyclic loading results, the primary outcomes were the construct axial and rotational displacements, measured by the MTS machine, at the end of each stage of loading. For each outcome, a linear mixed-effects model with Bonferroni correction was used to compare the two screw types, with loading stage as a repeated factor ^43,44,47. Survival analysis for specimen failures across all loading stages was performed using a log rank test for the groups at their corresponding loading stage at failure. For analysis of the four-point bending results, primary outcomes were the maximum force and stiffness. A Wilcoxon signed rank test was used for each outcome. Lastly for determining relationships between output variables and bone density, repeated measures correlation was used ⁴⁸.

RESULTS:

Compression Testing:

In tests of compression in the tibia specimens, interfragmentary compression forces were significantly higher in VPLS than in SLS after clamp removal (512 ± 324 N (standard deviation) vs. 79 ± 64 N, p =0.002) (Fig. 3A-B). In tests of plate-bone compression, the VPLS also produced forces that were significantly higher than SLS (242 ± 119 N vs. 104 ± 123 N, p=0.028) (Fig. 3C-D). The plate-bone compressive forces were greater in the VPLS in 79% of paired tests (11/14). There was no correlation found between bone mineral density and either interfragmentary or plate-bone compression (Fig. S-1).

Figure 3:

Compressive force mean values and individual specimen measurements after the clamp release for VPLS and SLS groups at each measurement stage recorded at the (A-B) interfragmentary interface and (C-D) plate-bone interface. P values indicate significant differences between screw types; only the final clamp release stage was used for statistical comparisons due to its clinical importance. Error bars indicate +/− standard error of the mean.

Stability Testing - Cyclic Loading:

In cyclic axial-torsional loading of the distal fibula plated constructs, all specimens from both screw type groups endured peak loads up to 80N and 2.5 N∙m. By the end of the final stage of loading to 180 N and 5 Nm, 3 specimens from each group had failed (Fig. 4C). All failures were due to fracture distal to the osteotomy site either below or between the two screws. No significant differences were detected between groups in the construct axial displacement or the construct rotation angle over all stages of loading (p > 0.05) (Fig. 4A-B). No significant differences were detected between groups in survival analysis across the loading stages (p = 0.827) (Fig. 4C). There were no significant differences in bone mineral density between external loading groups (p > 0.05).

Figure 4:

The average maximum (A) axial displacement and (B) rotation angle for each stage of cyclic combined loading. Data are plotted for the initial loading stages that were endured by all 7-matched pairs. (C) Survival plot showing fraction of samples that did not fail across all loading stages tested. Error bars indicate +/− standard error of the mean.

Stability Testing - Four-Point Bending:

In four-point bending tests of proximal fibula constructs, no significant differences were detected in either maximum force (SLS: 879.0 ± 461.7 N vs. VPLS: 782.1 ± 378.2 N, p = 0.44) or bending stiffness (SLS: 212.9 ± 58.4 N/mm vs. VPLS: 174.6 ± 40.7 N/mm, p = 0.44) (Fig. 5). Specimens failed either within the osteotomy or outside at the outermost screws.

Figure 5:

The group averages from 4-Point-Bend tests for (A) maximum force to yield and (B) construct stiffness. Error bars indicate +/− standard error of the mean.

DISCUSSION:

In tests of single screw compression in distal tibias, VPLS resulted in significantly larger mean compressive forces than SLS, at both the interfragmentary and plate-bone interfaces (Fig. 3). In tests of plated fibular constructs under cyclic axial-torsional loading, no differences in mean overall construct motions and failure were detected between the screw types (Fig. 4). In tests of plated fibular constructs in four-point bending, no differences in construct strength or stiffness were detected.

The interfragmentary compression provided by VPLS could be beneficial for primary bone healing in certain locking plate constructs, theoretically closing small fracture gaps and providing greater resistance to interfragmentary micromotion under external physiologic loads. The compression magnitudes were variable across specimens (Fig. 3B,D). No relationship was found between compression force and bone mineral density, with our density calculation considering both cortical thickness and local Hounsfield brightnesses. The mean compressive forces (in tibiae) were higher than reported in previous studies of variable pitch screws, which can be attributed to this study using bicortical screw purchase, larger screws, and clamping during screw insertion.^35-37

There were also increases in plate-bone compression with the VPLS compared to the SLS, although less pronounced than the differences in interfragmentary compression. This was likely due to the shorter span between the engaging screw head in the plate and the threads in the near cortex, providing fewer variable pitch threads to contribute to the compression, compared to the longer span between near and far cortices involved in interfragmentary compression ⁴⁹. Increased compression with the VPLS could indicate ability of this screw type to close small gaps between plate and bone during fixation, although this was not specifically tested. The compression present even with the SLS was at least partly due to the plate being clamped to the bone during screw insertion. To our knowledge there are no studies that measure plate-bone compression using a lag-screw to reduce a fracture through a plate. Plate-bone compression from non-locking screws (not spanning a fracture) has been reported in only a limited number of studies. Previous studies that plated cadaveric upper extremities using multiple non-locking screws measured higher compression forces with pressure films^50,51; however forces were due to a series of screws across a full construct, not one screw. A recent study reported a range of plate-bone compression forces using a single non-locking screw in several different model types⁵².

Despite the observed increased compression with VPLS in tibiae, differences in overall construct stability between the VPLS and SLS constructs were not detected in both the cyclic and four-point bending loading models of fibular constructs. The fracture fixation constructs involved a stable fracture pattern and bicortical locking fixation, resulting in relatively high resistance to external loading irrespective of screw type ^53,54. In our samples there was one additional failed specimen in the SLS group at several timepoints, but this did not raise to the level of statistical significance, and at the final timepoint there were an equal number of failures (3) in each group.

The simplified planar fracture models used in the compression tests were designed to enable accurate measurement of compression associated with insertion of a single screw. The posterior aspect of the distal metaphyseal tibia provided a relatively flat surface for a flat plate. A longitudinal osteotomy, separating the anterior and posterior cortices, enabled orientation of the load cell parallel to the plate and perpendicular to the screw axis. This geometry thus minimized load cell shear. The load cell’s through-hole design allowed the screw to be passed through its center, and load flanges resulted in all compressive loading from the screw to be transferred through the sensor. In pilot specimens, we attempted to use thin-film pressure sensors to measure compressive loads, but encountered substantial variability and inaccuracies associated with several factors.

Limitations of the compression testing included use of the simplified fracture model and effects of insertion of the load cell at the interfragmentary or plate-bone interface. Although a low-profile load cell was used, longer VPLS and SLS screws were needed to accommodate the load cell’s thickness. This may have affected the differential in thread pitch and resulting compression. Compression levels would also be affected by unicortical instead of bicortical fixation or testing at different anatomical locations having different screw length, cortical thickness, and bone densities. Additionally, the initial clamping force was deemed necessary and was determined through pilot testing, although this force is not measured clinically and may vary. Furthermore, each specimen underwent 4 compression tests using 4 different bone holes, with possible carryover effects, (although screw order was randomized). Cyclic axial-torsional and four-point bending tests were performed with constructs having stable fractures and bicortical fixation; testing with unstable fractures and/or unicortical fixation was beyond the present scope. Additionally, the findings of indifference in construct stability between VPLS and SLS may be limited to bicortical locking plate fixation of small bones such as the fibula. Future work could also compare the lag-by-technique use of a cortical screw to a VPLS screw in locking plate constructs.

Conclusions:

In this cadaveric biomechanical study, the variable pitch locking screws had greater interfragmentary and plate-bone compression when compared to standard locking screws in the context of insertion of single bicortical screws in the distal tibia. In the second part of the study, fibular constructs with stable fractures and bicortical fixation fixed with either screw type performed similarly under external loading. The use of variable pitch locking screws may be beneficial in locked constructs when obtaining interfragmentary compression (along the axis of the screw) is desired.