Stem size and stem alignment affects periprosthetic fracture risk and primary stability in cementless total hip arthroplasty

The ideal stem size and stem position is important for the success of total hip arthroplasty, since it can affect early implant loosening and periprosthetic fractures (PPF). This study aimed to investigate how small deviations from the ideal stem size and position influences the PPF risk and primary stability. Six experienced surgeons performed preoperative templating based on which the benchmark size for each femur was determined. Consecutive implantations were performed in six cadaveric femur pairs—one side was implanted with an undersized stem followed by the benchmark size and the contralateral side with a benchmark size followed by an oversized stem (Corail, Depuy Synthes). Moreover, three different alignments (six varus, six neutral, six valgus‐undersized) were compared using 18 femurs. Cortical strains during broaching and implantation were measured, and laser scans were used to determine final stem position. All specimens underwent dynamic loading. Primary stability was estimated from stem subsidence and pull‐out forces. Templated stem size varied between surgeons (±1 size; p = 0.005). Undersizing increased stem subsidence by 320% (p < 0.001). Oversized stems exhibited 52% higher pull‐out forces (p = 0.001) and 240% higher cortical strains (p = 0.056). Cortex strains increased with varus alignment (R2 = 0.356, p = 0.011) while primary stability decreased with valgus stem alignment (p = 0.043). Surgeons should be aware that small deviations from the ideal stem size and malalignments of the stem can significantly alter the mechanical situation and affect the success of their surgery.

1][12][13] In contrast, the question of whether undersizing reduces or promotes PPF risk is discussed controversially in the literature. 14,15Concerning oversized stems, the situation appears to be less controversial.Spina and Scalvi recognized that intraoperative PPF risk increases when an oversized stem is implanted 16 and Jasty and colleagues reported increasing cortical strains with increasing stem sizes. 17However, the degree to which the selected stem size can deviate from the "ideal" size before having a significant impact on primary stability and the risk of PPF is uncertain.
The position of the implanted stem is also dependent on processrelated factors such as the chosen surgical approach, the degree of exposure, and the level of surgical experience. 18[21] Most studies that address stem size and position are clinical follow-up studies.Radiographic lines on X-rays have been associated with stem loosening or revisions due to PPF with an additional evaluation of the relative stem size. 11,22The definition of over-or undersize of a stem and the extent to which it deviates from the apparent ideal stem size is not consistent. 11,13,16,23Nonclinical studies used bone replacement models 10 and numerical approaches. 12,21,24These approaches have limitations with regard to the accurate representation of the bone-implant interface.To address the clinical question of how the selected stem size and position affect the mechanical loading situation in the bone, it is important to investigate the impact of small deviations from the ideal stem size commonly encountered in clinical practice, 8 as well as the influence of misalignments in stem positioning using human samples.
The aim of this in vitro study was, therefore, to investigate how small deviations from the ideal stem size and variations in stem alignment influence PPF risk and the primary stability of the implant.

| METHODS
Cementless hip stems (Corail; Depuy Synthes) with a hydroxyapatite coating and standard offset of 135°were implanted in 15 femur pairs harvested from mostly male donors (m/f = 13/2) with an average body height of 175 ± 10 cm and age of 71 ± 17 years and stored below −20°C. 25The Ethics Commission of the Medical Association Hamburg (WF-067/18) approved this study.

| Preoperative planning
For the first part of the study, idealized two-dimensional scout views of six femoral pairs were reconstructed from CT scans and provided to six different surgeons to perform standard preoperative planning (Matlab 2020b; The MathWorks).The stem size for each femur was selected by each surgeon and the bone morphology determined based on Dorr types. 28The median templated stem size of the six surgeons for each femur was defined as the benchmark ("ideal") stem size.The remaining nine femoral pairs were templated by the single surgeon who performed the preparation and implantation in the second part of the study.The surgeons templating philosophy was in line with the benchmark size in the first group.

| Group 1: Stem size
The six femur pairs of the first group were used to evaluate the effect of stem size.Two consecutive stem implantations were performed in each femur.In one side of a pair, first an undersized stem of one size smaller than the benchmark size was implanted and analyzed, followed by a second implantation and analysis using the benchmark stem size (Figure 1A).On the contralateral sides, the benchmark stem size was implanted first, followed by an oversized stem of one size larger (Figure 1B).Femurs were randomly assigned to the respective groups.

| Group 2: Stem alignment
The stems in the second group were implanted in three deliberate alignments into nine femur pairs (six varus benchmark size, six neutral benchmark size, six valgus undersized).Excessive lateral opening of the canal was performed to allow valgus alignment of the stems, while lateral opening was omitted for the varus alignment cases (Figure 2).No specific target angles were defined prior femur preparation.To reflect the worst-case scenario in terms of primary stability, the combination of valgus and undersized stem was chosen.The other extreme-varus in combination with an oversized stem-was not achievable.
Soft tissues were removed from all specimens before they were visually aligned using the long femoral axis and embedded into pods (Technovit 4004; Kulzer GmbH).The specimens were mounted according to the surgeon's clinical approach (anterolateral approach).
The medial region of the calcar was carefully cleaned and degreased with acetone.Three millimeters below the head resection line, approximately 20-30 round markers (Ø: 0.4 mm) were randomly placed on the medial cortex of each femur in an area of about 25 × 50 mm for the determination of cortical strains using direct image correlation (DIC).
Preparation of the femur by compaction broaches and stem implantation was performed manually (weight of the mallet: 1.4 kg) with the corresponding instruments.Dynamic impaction forces during cavity preparation were measured using a piezoelectric load cell (Kistler 9333A) placed between the broach and the broach handle (Figure 3A).
Forces were recorded for the last two broaches (sample rate of cortical strain measurements using digital image correlation (DIC, 25 fps, 5 μm measurement noise; ARAMIS 3D Camera; GOM) with a measurement volume of 100 × 80 × 50 mm 3 (Figure 3B).
After implantation the stem tapers were cleaned and assembled with a ceramic ball head (BIOLOX ® delta; Depuy Synthes, Ø 36 mm, +5 mm).The position of the stem in the femur was acquired using laser scans (max.resolution 0.05 mm, max.accuracy 0.03 mm; HandySCAN 700, Creaform).A virtual 3D representation of the stem alignment in the femur was then reconstructed by superimposing the femoral cortices from laser scans and native qCTs (root mean square error below 0.3 mm).The canal-fill ratio (CFR) 7, vertical stem position, stem inclination, and cortical distance were evaluated using deterministic approaches to assess the size and alignment of the implanted stem relative to the femur.

| Mechanical testing
For mechanical testing the embedded specimens were aligned according to ISO 7206-4 with a 10°lateral and 9°dorsal tilt of the In one femur, a consecutive implantation was performed with an undersized and a benchmark stem (group 1A).The contralateral femur received a benchmark stem followed by an oversized stem (group 1B).Between the consecutive implantations, appropriate reaming was performed for the following stem.
(A) Preparation of the proximal femur for a targeted varus alignment of the stem.(B) Preparation of the proximal femur for a targeted valgus alignment of the stem.
The piezo load cell between the broach and the broach handle to measure forces during broaching and implantation.(B) Measurement of cortical strain using a stereo camera system and analysis based on digital image correlation.implant axis (2D spirit level) with respect to the loading axis and fixed in a ball and socket clamp (Figure 4A).Force-controlled cyclic loading was applied via a polyethylene covered piston to the ball head on the prosthesis using a servo-hydraulic testing machine (1 Hz; Bionix, MTS).Group 1 femora were cyclically tested using one loading condition (300 cycles; 80-800 N).Group 2 femora were loaded with two load levels (80-800 N; 80-1600 N), each applied for 600 cycles.Relative motion between the bone and the implant was measured contactless using DIC for the specimens of group 2. Recordings for 5 s were taken every 200 cycles (25 fps; measurement volume of 100 × 80 × 50 mm 3 , marker size 0.4 mm, ARAMIS 3D Camera, GOM, Figure 4B).Specimens were kept moist with ringer solution.After mechanical loading, an additional laser scan was performed to determine the final stem subsidence after loading.Stems were finally removed displacement-controlled with a speed of 0.1 mm/s and pull-out forces were recorded (Z010; Zwick Roell).

| Statistical analysis
Statistical analysis was performed with a type I error level of 0.05 (IBM SPSS Statistics 24; IBM Corp).Unless otherwise described, the mean and standard deviation are reported.Relationships between parameters were determined using multiple Pearson correlations.Repeated-measures analyses of variance were performed to test and evaluate the eligibility of the study design with repeated implantations.Differences within a femoral pair were analyzed using paired t tests.Differences between conditions were examined using analyses of variance.

| Group 1: Stem size
The preoperatively planned stem size for the six femur pairs of group 1 varied among the six surgeons by ±1 size (median stem size: 12 [size 11, size 13]; p = 0.005).Variations were also found in terms of the planned vertical stem position (9.7 ± 2.9 mm up to 14.6 ± 4.1 mm, p < 0.001).
The two consecutive stem implantations into one femur did not significantly affect the applied forces during broaching (p = 0.685; Figure 5A).Cortical strains, stem subsidence and pull-out forces were also similar in both groups for the benchmark stem sizes (p = 0.746, p = 0.403, p = 0.077; Figure 5B,C) allowing pooling of the data for further analysis.had no effect on the stem subsidence.Oversized stems showed higher pull-out forces compared to benchmark stems (p = 0.008), while implantation of undersized stems resulted in lower pull-out forces (p = 0.007; Figure 7A).Concurrently, femoral morphology influenced pullout forces for all stem sizes showing a decrease in force with increasing CCS (oversized: R 2 = 0.740, p = 0.028; benchmark: R 2 = 0.700, p = 0.001; undersized: R 2 = 0.546, p = 0.093, 1 − β = 0.3745; Figure 7B).

| DISCUSSION
The six experienced surgeons exhibited a variation of ±1 stem size during templating.In in vitro experiments, this small variation lead to stem subsidence in case of undersized stems and to higher cortical strains for oversized stems or varus alignment.
The femur pairs used for stem implantations showed little variations with respect to bone quality and bone morphology and no outliers with for example, like osteoporotic bone, which is more common in elderly females. 6The specimens such represent a standard patient collective receiving uncemented primary THA.

| Methodology and limitation
Defining the benchmark size on the planning of several surgeons is regarded as more representative for the clinical situation than a size based on the choice of a single surgeon. 11,12,17,24udies using paired femurs reduce the effects of patient-specific factors for the comparison between two levels of an independent factor.In this study, a new approach was introduced to even further reduce the influence of patient specific factors by performing consecutive implantations in the same femur.This approach additionally has the economic effect that fewer human specimens are needed.
It could be shown that this proceeding does not affect the interface between stem and bone if broaching for one size larger is performed.
This supports the existing clinical practice of implanting the next stem size when intraoperative primary stability is not achieved with the selected stem size.Previous studies using consecutive implantations have been performed without validation. 17,29e to limited availability, cadaveric studies have to deal with small sample sized and low power of the experiment.Further it is limited to the single type of implant which is a fully porous HAcoated tapered wedge stem (Corail) and provides metaphyseal and diaphyseal contact.Undersized valgus stems were compared to bench-mark sized implants in different stem orientations but not to varus oversized, since the surgeon was not able to implant them.

| PPF risk
The choice of stem size influences the periprosthetic fracture risk.
Higher medial cortical stresses with implantations of oversized stems have already been reported 16,17 and even slightly oversized stems (by one size) result in a larger amount of removed bone.Additionally, larger preparation broaches are related to higher preparation forces and cortical strains. 30The magnitude of the local strain peaks of 5200 με measured during the implantation of oversized stems are close to the critical fracture range of 5500-10000 με. 17,31In contrast to Bonnin and colleagues who indicated an increased risk of PPF when oversized stems were implanted in small a.p. diameter femora, 32 no increase in cortical strains associated with a Dorr Type A femoral morphology (small CCS) was observed in this study.
Implantation of an oversized stem into a Dorr Type C femur with an aged bone morphology could increase the risk of PPF due to the higher broaching and impaction forces combined with a reduced distance to thin, fragile cortical bone, as cautioned already in the literature. 15em alignment is shown to affect strain and such the PPF risk.
[21] The observed mean cortical strains between 3200 and 4300 με were associated with local strain peaks of 8000 με, which lie within the critical range for fracture initiation between 5500 and 10,000 με. 17,31e reduced distance between the stem and the cortex due to varus stem alignment inevitably leads to a decreased thickness of the trabecular layer between stem and cortex.This reduces the forcedistributing and dampening effect of the trabecular layer and might promote an uneven load distribution with higher local loads on the cortex. 20In addition to the altered loading situation of the bone, incorrect stem alignment also affects the reconstruction of the head center of rotation. 18

| Primary stability
The relationship between the primary stability of the stem and bone quality or morphology has been established in previous studies, 33,34 the present findings particularly emphasize the influence of stem size on primary stability.The higher primary stability of oversized stems can be attributed to the higher degree of canal filling ratio in the proximal femur and tighter fit of the stem.Vice versa the highest stem subsidence (~0.8 mm) was found for undersized stems when loaded with single body weight (800 N).Higher loads may further increase stem subsidence and might cause critical subsidence during the initial phase after implantation preventing ingrowth (1.2-3 mm 35,36 ) or the initiation of a PPF due to excessive stem subsidence. 14The lower pullout forces for undersized stems are in line with observations from previous studies and further underline the importance to avoid undersizing. 13,23,33,35Primary stability as assessed in this study was improved for stems with a varus malalignment, but the higher relative motions between bone and stem also observed may negatively influence bone ingrowth.The lower primary stability observed for valgus undersized stems cannot be clearly attributed to the valgus alignment, as it could just as well have been caused by the smaller stem size.

| CONCLUSION
Surgeons should be aware that for fully porous HA-Coated tapered wedge stems such as the Corail even minor deviations from the ideal stem size, can significantly alter the mechanical situation and may affect primary stability or PPF risk and such the success of the own surgery.Malalignment of the stem, especially into varus direction, was shown to significantly increase cortical loading and thus the PPF risk in addition to altering the physiologic reconstruction of the joint center.
Cortical strains were similar during broaching and stem implantation (broaching: 580 ± 1090 με, implantation: 580 ± 810 με, p = 0.979).Higher cortical strains were observed with larger relative stem sizes (CFR; R 2 = 0.267, p = 0.010).The highest cortical strains on the medial side were measured for oversized stems (p = 0.056, local peak strains up to 5.200 με; Figure6A).Undersized stems showed more stem subsidence than benchmark stems or oversized stems (p = 0.001; Figure6B).The femoral morphologyF I G U R E 4(A) 2D spirit level was used for precise alignment of the implanted stem according to ISO 7206-4 (10°lateral inclination, 9°dorsal inclination) and fixation via the ball socket holder.(B) Cyclic loading was applied to the ceramic head via a polyethylene piston.Relative motion at the bone-implant interface was measured contactless using digital image correlation.F I G U R E 5 Broaching forces (A), cortical strain (B), stem subsidence (C), and pull-out forces (D) for initial and consecutive stem implantations (shown are means and standard deviations).

F I G U R E 6 (
A) Cortical strain measured at the proximal medial cortex for the three stem size groups.(B) Stem subsidence during mechanical loading with respect to the stem size groups (shown are mean and standard deviation).

F
I G U R E 7 (A) Pull-out force in relation to the implanted stem size (mean and standard deviation).(B) Pullout forces decrease with increasing CCS for the three different stem size groups.F I G U R E 8 (A) Influence of stem alignment on the cortical strains on the medial proximal cortex showing an increase with a change from valgus to varus.(B) Influence of the medial cortical distance on the cortical strains on the medial proximal cortex showing a decrease in strain with a change from varus to neutral and valgus.