Biomechanical Evaluation of Various Screw Constructs for Proximal Phalanx Base Fracture Fixation

Abstract

Purpose

To evaluate the biomechanical properties of four fixation methods for proximal phalanx base fractures, single antegrade cannulated screw (control), V-pattern, Y-pattern, and K-wire configurations, using a matched pair cadaveric model.

Methods

Forty-eight fingers (index through little) from six matched cadaveric hand pairs were included. A standardized transverse base fracture was created in the proximal phalanx. One side of each pair received a single antegrade cannulated screw. Contralateral digits were treated with either a V-pattern (group 1), Y-pattern (group 2), or K-wire configuration (group 3). Specimens underwent 3-point bending, rotational testing, and cantilever bending to failure using a hydraulic test frame. Key metrics included bending stiffness, torsional stiffness, axial displacement, and peak load to failure.

Results

The V-pattern demonstrated considerably higher bending stiffness, reduced displacement, greater rotational stiffness, and higher torque resistance compared to controls. While failure displacement was similar, the V-pattern required considerably higher force to fail and showed greater stiffness. The Y-pattern showed no notable differences in bending stiffness or displacement versus controls. However, it had considerably lower rotational stiffness and torque resistance. No differences in failure load, stiffness, or displacement were noted. The K-wire showed considerably lower bending stiffness and higher displacement compared to controls. Rotational metrics and failure parameters showed no notable differences.

Conclusions

Among the configurations tested, the V-pattern demonstrated superior biomechanical stability across bending, torsional, and failure testing compared to single antegrade cannulated screw and dual K-wire supporting our hypothesis. These findings support further investigation into its clinical application for improved fixation in proximal phalanx base fractures. Moreover, the addition of a smaller screw to make the Y-pattern did not add considerable strength.

Type of study/level of evidence

Diagnostic IV.

Hand fractures in adults are a common occurrence, accounting for 19% of all fractures, where 59% of these hand fractures were of the phalanges and 38% of those were comprised of proximal phalanx fractures.¹ Although a variety of fracture types can occur, a simple transverse base fracture is commonly seen, especially in older patients.² Stable proximal phalanx fractures can be treated nonoperatively with splinting, as a gentle range of motion does not result in displacement; however, unstable proximal phalanx fractures either displace with range of motion or have unsuccessful fracture reduction and require operative intervention.³ Proximal phalanx fractures experience deforming forces from both the intrinsic and extrinsic muscular system, and therefore are inherently unstable. Apex volar deformity presents namely as flexion from the interossei muscle insertion on the base of the proximal fragment and extension from the central slip on the middle phalanx base.^2,3 The extrinsic flexor and extensor tendons shorten the phalanx and extend the distal fragment via their compressive forces.³

The use of K-wire is a common fixation method with good results being reported,^4,5 but a complication rate of up to 36% has been reported.⁶ Moreover, because it requires a period of postoperative immobilization, joint stiffness is not uncommon.^7,8 During fracture fixation, protection of the soft tissue envelope is paramount for optimal functional outcomes. Alternative operative interventions have been used and studied, such as rigid fixation with lag screws, plates, or both, and use of cannulated screws/nails. The advantages of cannulated screws include immediate range of motion, decreased scar burden and thus decreased adhesions, and avoidance of periosteal stripping.³ Several biomechanical studies have tested the various fixation methods for proximal phalanx fractures in cadaver specimens, including K-wire fixation, headless compress screw fixation, and plate and screw constructs.^2,9, 10, 11, 12 There is currently a paucity of data evaluating the performance of different dual intramedullary headless compression screws. Gaspar et al¹³ adopted the dual screw Y-strut described by del Piñal et al¹⁴ in metacarpal fractures and reported good clinical results with 1-year follow-up when treating proximal phalanx fractures. Conversely, Lie et al¹⁵ investigated the same Y-strut configuration comparing it against the single intramedullary screw. They concluded that there was no considerable difference in stiffness, and the Y-strut resulted in lower load to failure. While there are studies comparing headless compression screw fixation to other methods of fixation, to the best of our knowledge, there are no studies to date comparing the single antegrade cannulated nail to multiple screw constructs such as the V- and Y-pattern configurations for the management of proximal phalanx base fractures.

The purpose of this study was to evaluate the biomechanical properties (torsional stiffness, bending stiffness, and load to failure) of four different repair configurations for proximal phalanx fractures: V-pattern, Y-pattern, K-wire fixation, and single antegrade cannulated screw. It was hypothesized that the V-pattern will provide a considerable advantage in biomechanical properties, leading to improved fixation, when compared to the Y-pattern, K-wires, and single antegrade cannulated screw configuration for proximal phalanx fracture repairs.

Materials and Methods

Sample grouping

This was a controlled biomechanical study. The proximal phalanges of the index, middle, ring, and little fingers from six matched pair cadaveric hands (totaling to 48 fingers) were divided into the following testing groups: group 1—The control samples received the single antegrade cannulated nails as treatment while the contralateral side (experimental samples) received the V-pattern configuration; group 2—The control samples received the single antegrade cannulated nails as treatment. The contralateral side (experimental samples) received the Y-pattern configuration, where the addition of the smaller screw may act as a “kickstand”; group 3—The control samples received the single antegrade cannulated nail as treatment, and the contralateral side (experimental samples) received the K-wire configuration, which crosses as an X-pattern.

Within pairs, the right and left fingers were randomized to address any effects of hand dominance and possible differences in bone quality. All samples were instrumented with InFrame micronail (2.0 mm, Exsomed) varying in length depending on fixation method.

Sample preparation: fracture creation and instrumentation

All bones were investigated for prior pathology by fluoroscopic radiography. The index, middle, ring, and little fingers of all hands were harvested and stripped of all soft tissue, resulting in a total of 48 available samples. The proximal phalanx was then isolated and disarticulated from the metacarpophalangeal and the proximal interphalangeal joints.

The fracture pattern chosen simulated a proximal phalanx base fracture. The location of fracture was standardized to be 20% of the phalanx length, representative of the flare junction. With the aid of fluoroscopy, a guidewire was initially inserted to the bone for positioning and creating a pilot hole for each grouping. The desired implant length was determined by the surgeon intraoperatively based on length markings at the guidewire shoulder (diameter transition point). Thereafter, a partial transverse fracture in the metaphyseal region was created using an oscillating bone histology saw and all were performed by senior author. The appropriate screw configuration was then instrumented accordingly to maintain alignment. Once the implant was set, the fracture was completed taking care not to contact implant (Fig. 1).

Figure 1.

Examples of each construct, where the top row is the V-pattern, the Y-pattern, and the K-wires. Bottom row is the single antegrade cannulated screw, control.

Biomechanical testing

Each sample was mechanically tested in the following order: 3-point bending, rotational testing, and cantilever bending load to failure. During the 3-point bending (Fig. 2A), each sample was placed horizontally onto platforms coupled to the test frame. The distance of the lower support platform was kept at 20 mm, and the loading support was at the midpoint and 1 cm from the fracture line.⁸ Each sample was then cyclically loaded from 10 to 50 N for 25 cycles at 0.5 Hz. Thereafter, each phalanx was embedded with high-strength resin (Bondo, 3M) in a custom fixture mounted vertically to the hydraulic test frame (MTS Bionix; MTS Systems). The distance from the potting to the fracture line was 2 cm. The axis of rotation was along the center line of phalanx allowing torsional rotation of the head of phalanx (Fig. 2B). Torsional rotation was applied with 25 cycles from 1° to 2° at 0.5 Hz. The direction of rotation will be standardized to counterclockwise for all samples. The specimen was placed in a horizontal position for axial loading perpendicular to the dorsal surface of the specimen and was applied for ramp to failure testing (cantilever bending; Fig. 2C). The loading pin was centered over the articular surface in a dorsopalmar direction at a distance of 3 mm from the fracture. Force was applied at a rate of 100 mm/min until failure. All samples were subject to the same loading sequence.

Figure 2.

Illustration of biomechanical testing: A 3-point bending, B torsional testing, and C cantilever bending to failure.

The metrics of interest were collected from the test frame data acquisition system. The load-displacement curve generated was used to calculate bending stiffness (N/mm), torsional stiffness (N mm/°), axial displacements (mm), and peak load to failure (N). Stiffness values corresponded to the linear slope for each load-displacement curve (calculated from average of the last five cycles), whereas the peak load to failure will be defined as an abrupt change in the load-displacement curve.

Statistical analysis

Based on data from literature, an a priori power analysis was used with an effect size (Cohen’s d) of 1.4 to compute the required sample size to detect a difference between the matched paired groups (G∗Power V3.1.9.2; Franz Faul).^2,11,12,16 Assuming a power of 0.8 and a type I error rate of 0.05, a sample size of seven per group powers the study to 0.86. The choice of eight samples per group powers the study to 0.92 allowing for any possible tissue rejection, or unforeseen failures while retaining proper study power. Commercially available software will be used for all comparisons (SPSS) at a significance threshold of 0.05. Because of the smaller sample size and non-normally distributed data, nonparametric statistical analysis was performed using Kruskal-Wallis test for each group. Data were presented as mean ± SD.

Results

Three-point bending

In G1, the V-pattern configuration was found to have significantly higher stiffness when compared to the control group (405.76 ± 123.5 N/mm vs 213.22 ± 50.7 N/mm; P = .003) and had significantly less displacement (0.19 ± 0.1 mm vs 0.08 ± 0.013 mm; P = .002) during 3-point bending. One sample from group 1 control was excluded as it was highly unstable, and the matched pair in the V-pattern was removed from comparison, resulting in a sample size of seven for G1.

In G2, the Y-pattern configuration was found to have no considerable differences when comparing both stiffness (250.4 ± 89.6 N/mm vs 268.7 ± 100.5 N/mm) and displacement to the control group during 3-point bending (0.16 ± 0.1 mm vs 0.20 ± 0.2 mm).

In G3, the K-wire configuration was found to have significantly less stiffness when compared to the control group (122.8 ± 63.7 N/mm vs 203.49 ± 59.7 N/mm; P = .036) and had significantly more displacement (0.28 ± 0.1 mm vs 0.15 ± 0.05 mm; P = .027). Data are summarized in the Table 1.

Table 1.

Data Summary of Stiffness and Displacement During 3-Point Bending^∗

Study Group	n	Stiffness (N/mm)	Axial Displacement (mm)
Control	7	213.2 ± 50.7 (166.3−260.1)	0.19 ± 0.054 (0.14−0.24)
V-pattern	7	405.8 ± 123.5 (291.6−520.0)	0.085 ± 0.026 (0.060−0.11)
P		.003	.002
Control	8	268.4 ± 93.0 (190.6−346.2)	0.20 ± 0.16 (0.068−0.33)
Y-pattern	8	250.4 ± 89.6 (175.5−325.3)	0.16 ± 0.083 (0.089−0.23)
P		.6	.916
Control	8	203.5 ± 59.7 (153.6−253.4)	0.15 ± 0.043 (0.12−0.19)
K-wire	8	122.8 ± 63.7 (69.6−176.1)	0.28 ± 0.11 (0.19−0.37)
P		.036	.027

^∗Data are presented as mean ± SD (95% CIs).

Rotational testing

In G1, the V-pattern configuration was found to be more rotationally stable when compared to the control group. The rotational stiffness was significantly larger (124.5 ± 36.5 N mm/° vs 63.6 ± 56.3 N mm/°; P = .035), and the resistance to rotation (torque) was significantly higher (278.9 ± 78.9 N mm vs 107.2 ± 63 N mm; P = .006).

For G2, the Y-pattern configuration was found to be less rotationally stable when compared to the control group. The stiffness was significantly less (56.3 ± 48.3 N mm/° vs 144.3 ± 57.1 N mm/°; P = .006), and the torque was significantly lower (154.5 ± 61.4 N mm vs 241.5 ± 90.2 N mm; P = .027).

In G3, no significant differences in rotational stability (97.3 ± 68.6 N mm/° vs 45.8 ± 37.9 N mm/°) or amount of torque (152.1 ± 85.1 N mm vs 128 ± 48.4 N mm) were observed between the K-wire configuration and the control group. Data are summarized in the Table 2.

Table 2.

Data Summary of Torsional Stiffness and Torsion During Rotational Testing^∗

Study Group	n	Torsional Stiffness (N mm/°)	Torsion (N mm)
Control	7	63.6 ± 56.3 (11.5−115.7)	107.2 ± 63.0 (49.0−165.5)
V-pattern	7	124.5 ± 36.5 (90.7−158.2)	278.9 ± 79.0 (205.9−351.9)
P		.035	.006
Control	8	144.3 ± 57.1 (96.6−192.0)	241.5 ± 90.2 (166.0−316.8)
Y-pattern	8	56.3 ± 48.3 (15.9−96.6)	154.5 ± 61.4 (103.2−205.8)
P		.006	.027
Control	8	97.3 ± 68.6 (40.0−154.6)	152.1 ± 85.1 (81.0−223.2)
K-wire	8	45.8 ± 37.9 (14.1−77.5)	128.0 ± 48.4 (87.5−168.5)
P		.059	.6

^∗Data are presented as mean ± SD (95% CIs).

Cantilever bending (ramp to failure)

During the ramp to failure in G1, the V-pattern and control failed at similar displacements. However, the V-pattern required significantly more force (67.3 ± 26.6 N vs 30.7 ± 13.8 N; P = .013) and was stiffer compared to the control (16.1 ± 9.1 N/mm vs 7.5 ± 3.8 N/mm; P = .035).

During the ramp to failure for G2, no significant differences were observed between the Y-pattern and control group for stiffness, failure load, or displacement.

Similarly, no significant differences were observed between the K-wire and control group for stiffness or displacement. However, the K-wire failed at a higher load compared to the control (57.1 ± 16.8 N vs 33.8 ± 18.6 N; P = .035). Additionally, one sample in G3 control dissociated while setting up for the test, where the proximal fragment separated. This sample and its matched pair were removed, resulting in a sample size of seven. Data are summarized in the Table 3.

Table 3.

Data Summary of Stiffness, Load to Failure, and Displacement During Ramp to Failure^∗

Study Group	n	Stiffness (N/mm)	Failure Load (N)	Axial Displacement (mm)
Control	7	7.5 ± 3.8 (4.0−11.0)	30.7 ± 13.8 (17.9−43.5)	5.8 ± 1.9 (3.9−7.6)
V-pattern	7	16.1 ± 9.1 (7.7−24.6)	67.3 ± 26.6 (42.7−91.9)	7.7 ± 2.8 (5.1−10.3)
P		.035	.013	.18
Control	8	13.5 ± 3.6 (10.4−16.5)	42.1 ± 18.2 (26.9−57.3)	5.8 ± 1.6 (4.5−7.1)
Y-pattern	8	17.3 ± 8.6 (10.1−24.5)	63.9 ± 33.5 (36.0−92.0)	5.6 ± 1.3 (4.6−6.7)
P		.345	.115	.753
Control	7	17.0 ± 10.6 (7.2−26.8)	33.8 ± 18.6 (16.6−51.0)	4.6 ± 3.0 (1.8−7.4)
K-wire	7	15.0 ± 6.6 (8.9−21.1)	57.1 ± 16.8 (41.5−72.7)	7.1 ± 1.3 (5.8−8.3)
P		.949	.035	.225

^∗Data are presented as mean ± SD (95% CIs).

Failure mechanism

All samples experienced the proximal fragment displacing resulting in bony cut-out. However, it was observed that the plane and amount of rotation the fragment experienced varied based on configuration. Rotations were estimated using Image-J referencing the images (before and after) ramp to failure testing (Table 4).

Table 4.

Data Summary of Fragment Rotation at Failure

Group	Coronal Plane Rotation (°)
G1_CTRL	62.7 ± 10.5
G2_CTRL	54.9 ± 10.6
G3_CTRL	71.4 ± 36.4
G2_YPattern	51.4 ± 7.8
Group	Sagittal Plane Rotation (°)
G1_VPattern	126.5 ± 9.9
G3_KWire	139.8 ± 11.3

The control groups experienced the most fragment rotation in the coronal plane, followed by the Y-pattern. This is likely because of the single screw behaves as a pivot point. The addition of the second screw in Y-pattern reduced the amount of rotation experienced. However, there was little deflection or rotation in the sagittal plane. Conversely, both the V-pattern and K-wire groups experienced fragment displacement in a linear fashion resulting in posterior cortical fracture and rotation in the sagittal plane (Fig. 3).

Figure 3.

Examples of failure modes where the control groups A experienced the most fragment rotation in the coronal plane, followed by the Y-pattern B. This is likely because of the single screw behaving as a pivot point. The addition of the second screw in Y-pattern reduced the amount of rotation experienced. However, there was little deflection or rotation in the sagittal plane. Conversely, both the K-wire (C, left) and V-pattern (C, right) groups experienced fragment displacement in a linear fashion, resulting in posterior cortical fracture and rotation in the sagittal plane D.

Discussion

The biomechanical performance provided by the V-pattern fixation was determined to be a considerable improvement compared to a single antegrade cannulated screw across all loading conditions. In addition, the V-pattern and K-wire design had similar rotation and fracture complications as a result of the displacement of the construct. Therefore, the use of V-pattern for proximal phalanx fracture fixation may provide improved bending and rotational stability, as well as increased failure loads. Clinically, this may allow for improved fixation. The addition of the smaller “kickstand” screw in Y-pattern does not appear to considerably impact the construct strength. A single screw placed into the far cortex is similar in strength to two crossing K-wires. These findings partially support our hypothesis that the V-pattern would provide a notable advantage in biomechanical properties compared to the other fixation methods.

Proximal phalanx fractures occur across most age ranges, making this study impactful by addressing the biomechanical properties of different constructs.^1,17 Previous studies have been performed using oblique fracture patterns, screws versus plates, and many other variables.^9,11,12 There is limited literature comparing different arrangements of cannulated screws with K-wires in this anatomical region. The proximal phalanx base articulates at the metacarpophalangeal joint, which has a higher degree of freedom than interphalangeal joints; this creates a need to report the biomechanical difference of each construct when a fracture occurs.^3,18

Gaspar et al¹³ reported successful results in a pilot series of patients using a technique described by del Piñal et al¹⁴ in metacarpal fractures as Y-strutting (Y-pattern). In their 1-year follow-up, it was reported that patients had excellent postoperative motion, near-normal grip strength, and no complications. While one cannot directly compare between a clinical case series and biomechanical investigation, our study found the Y-pattern configuration was found to be less rotationally stable compared to the control, despite having similar performance during 3-point bending and load to failure. This could be because of multiple factors, such as bone quality, cortical engagement of the screws, and depth of “kickstand” screw. While the specimens were matched pairs for comparison, it was observed that the quality of bone varied across groups. Some samples had large canals with little trabecular bone and thinning cortices, which may play a role in the screw engagement. All control samples and Y-pattern constructs had the “long” screw bicortical. Finally, the depth of each screw was crucial as the head of screw was to be below articular surface; however, the poor bone quality affected the fixation within the phalanx proximal fragment. Given the “kickstand” screw in the Y-pattern configuration was smaller, with less points of engagement in the bone, it did not appear to provide any biomechanical advantage and observed pulling out of the bone easily during failure testing. Lie et al¹⁵ also evaluated the biomechanical properties of this configuration and found that it too underperformed compared to a 3-mm single intramedullary screw resulting in lower failure loads. The second screw in this configuration is placed in an oblique orientation and therefore shorter; this may not sufficiently contribute to canal fill at the isthmus, which has been related to stability.^13,15,19 This can indicate the “kickstand” screw may not provide as much additional benefits, especially in poor bone quality. However, during failure testing, our results show it did reduce the amount of rotation experienced by the fragment compared to the control. Further investigations are warranted to see the effect of this reduced rotation and the outcomes of clinical series.

The opposite was observed for the K-wire group; rotationally, the performance was similar to the control, but in 3-point bending, the K-wire was less stiff and allowed more displacement. Our results show that a single screw is as strong as two crossing K-wires and this could be beneficial in clinical practice as it would allow for earlier range of motion and eliminate the need for wire removal later. We thus would recommend that if a single screw were used that it be placed into the far cortex and buried just subchondral on proximal aspect. When performing the V-pattern, it is recommended to start as radial/ulnar as possible to place the two screws. This configuration often will end in a V-pattern; however, based on placement, it may be more of an X configuration. Our results show that these two screws do not need to be placed into the far cortex.

This study was not exempt from limitations. This is a biomechanical study and has inherent limitation in correlation with clinical practice. Future clinical studies comparing outcomes of these various techniques could be performed to further understand differences more. This study also used cadaveric specimens and bone quality of each specimen must be taken into account. While the selection of matched pair samples helps mitigate these effects, bone quality could affect failure of constructs. Additionally, only one screw/nail company was used and there may be differences between various devices used in clinical practice. A comparison study between different diameter screws/nails may yield more insight. Moreover, these results only represent transverse fractures at the base of the proximal phalanx. Although this approach allowed standardization for mechanical testing, results should not be extrapolated to other fracture patterns as further investigation in complex patterns is required.

Conflicts of Interest

Dr Doarn has association or financial involvement with Axogen. Dr Collins has association or financial involvement with Acumed LLC and Endo. No benefits in any form have been received or will be received by the other authors related directly to this article.