Biomechanical properties of artificial bones made by Sawbones: A review

Biomedical engineers and physicists frequently use human or animal bone for orthopaedic biomechanics research because they are excellent approximations of living bone. But, there are drawbacks to biological bone, like degradation over time, ethical concerns, high financial costs, inter-specimen variability, storage requirements, supplier sourcing, transportation rules, etc. Consequently, since the late 1980s, the Sawbones® company has been one of the world's largest suppliers of artificial bones for biomechanical testing that counteract many disadvantages of biological bone. There have been many published reports using these bone analogs for research on joint replacement, bone fracture fixation, spine surgery, etc. But, there exists no prior review paper on these artificial bones that gives a comprehensive and in-depth look at the numerical data of interest to biomedical engineers and physicists. Thus, this paper critically reviews 25 years of English-language studies on the biomechanical properties of these artificial bones that (a) characterized unknown or unreported values, (b) validated them against biological bone, and/or (c) optimized different design parameters. This survey of data, advantages, disadvantages, and knowledge gaps will hopefully be useful to biomedical engineers and physicists in developing mechanical testing protocols and computational finite element models.

Biomechanical design optimization of distal femur locked plates: A review

Clinical findings, manufacturer instructions, and surgeon's preferences often dictate the implantation of distal femur locked plates (DFLPs), but healing problems and implant failures still persist. Also, most biomechanical researchers compare a particular DFLP configuration to implants like plates and nails. However, this begs the question: Is this specific DFLP configuration biomechanically optimal to encourage early callus formation, reduce bone and implant failure, and minimize bone "stress shielding"? Consequently, it is crucial to optimize, or characterize, the biomechanical performance (stiffness, strength, fracture micro-motion, bone stress, plate stress) of DFLPs influenced by plate variables (geometry, position, material) and screw variables (distribution, size, number, angle, material). Thus, this article reviews 20 years of biomechanical design optimization studies on DFLPs. As such, Google Scholar and PubMed websites were searched for articles in English published since 2000 using the terms "distal femur plates" or "supracondylar femur plates" plus "biomechanics/biomechanical" and "locked/locking," followed by searching article reference lists. Key numerical outcomes and common trends were identified, such as: (a) plate cross-sectional area moment of inertia can be enlarged to lower plate stress at the fracture; (b) plate material has a larger influence on plate stress than plate thickness, buttress screws, and inserts for empty plate holes; (c) screw distribution has a major influence on fracture micro-motion, etc. Recommendations for future work and clinical implications are then provided, such as: (a) simultaneously optimizing fracture micro-motion for early healing, reducing bone and implant stresses to prevent re-injury, lowering "stress shielding" to avoid bone resorption, and ensuring adequate fatigue life; (b) examining alternate non-metallic materials for plates and screws; (c) assessing the influence of condylar screw number, distribution, and angulation, etc. This information can benefit biomedical engineers in designing or evaluating DFLPs, as well as orthopedic surgeons in choosing the best DFLPs for their patients.

The biomechanical behavior of 3D printed human femoral bones based on generic and patient-specific geometries

Background

Bone is a highly complex composite material which makes it hard to find appropriate artificial surrogates for patient-specific biomechanical testing. Despite various options of commercially available bones with generic geometries, these are either biomechanically not very realistic or rather expensive.

Methods

In this work, additive manufacturing was used for the fabrication of artificial femoral bones. These were based on CT images of four different commercially available femoral bone surrogates and three human bones with varying bone density. The models were 3D printed using a low-budget fused deposition modeling (FDM) 3D printer and PLA filament. The infill density was mechanically calibrated and varying cortical thickness was used. Compression tests of proximal femora simulating stance were performed and the biomechanical behavior concerning ultimate force, spring stiffness, and fracture pattern were evaluated as well as compared to the results of commercial and cadaveric bones.

Results

Regarding the ultimate forces and spring stiffness, the 3D printed analogs showed mechanical behavior closer to their real counterparts than the commercially available polyurethan-based surrogates. Furthermore, the increase in ultimate force with increasing bone density observed in human femoral bones could be reproduced well. Also, the fracture patterns observed match well with fracture patterns observed in human hip injuries.

Conclusion

Consequently, the methods presented here show to be a promising alternative for artificial generic surrogates concerning femoral strength testing. The manufacturing is straightforward, cheap, and patient-specific geometries are possible.

Novel slide compression anatomic plates of the femoral neck for treating unstable femoral neck fracture: A biomechanical study

To compare the biomechanical stability of slide compression anatomic plates of the femoral neck, cannulated compression screws and dynamic hip screws with derotation screws for stabilizing unstable femoral neck fractures (Pauwels angle = 70°). Pauwels III femoral neck fractures were created on 45 Sawbones femurs and randomly assigned to three implant groups (1:1:1). The biomechanical stability of all Sawbones in each treatment group was evaluated with three tests. First, in the static loading test, the load-displacement curve, vertical stiffness (load/vertical displacement [N/mm]) and 5 mm failure load were recorded. Second, in the incremental cyclic loading test (700, 1000, and 1400 N), the cyclic-displacement curve and the displacement of the fragments were recorded. Third, in the torsion test, the torsional rigidity, maximum torque, and torsional angle corresponding to the maximum torque were recorded. The static compression test showed that slide compression anatomic place-femoral neck (SCAP-FN) had the largest vertical stiffness (275 ± 11 N/mm, p < 0.01) and 5 mm failure load (1232 ± 156, p < 0.01). The cyclic loading test showed that SCAP-FN had the lowest change in displacement after 30000 cycles of loading. The torsional stiffness and the maximum torque followed the order SCAP-FN > dynamic hip screw systems (DHS) + derotational screw (DS) > CCS, and the torsional angle corresponding to the maximum torque followed the order SCAP-FN < DHS + DS < CCS. The SCAP-FN construct provides stiffness and stability compared with other standard fixation techniques (3CS and DHS + DS). The fixation strategy of SCAP-FN might be sufficient for clinical use, indicating studies in the human body are warranted.

Biomechanical design using in-vitro finite element modeling of distal femur fracture plates made from semi-rigid materials versus traditional metals for post-operative toe-touch weight-bearing

This proof-of-concept study designs distal femur fracture plates from semi-rigid materials vs. traditional metals for toe-touch weight-bearing recommended to patients immediately after surgery. The two-fold goal was to (a) reduce stress shielding (SS) by increasing cortical bone stress thereby reducing the risk of bone absorption and plate loosening, and (b) reduce delayed healing (DH) via early callus formation by optimizing axial interfragmentary motion (AIM). Finite element analysis was used to design semi-rigid plates whose elastic moduli E ensured plates permitted AIM of 0.2 - 1 mm for early callus formation. A low hip joint force of 700 N (i.e. 100% x body weight) was applied, which corresponds to a typical 140 N toe-touch foot-to-ground force (i.e. 20% x body weight) recommended to patients after surgery. Analysis was done using 2 screw materials (steel or titanium) and types (locked or non-locked). Steel and titanium plates were also analyzed. Semi-rigid plates (vs. metal plates) had lower overall femur/plate construct stiffnesses of 508 - 1482 N/mm, higher cortical bone stresses under the plate by 2.02x - 3.27x thereby reducing SS, and lower E values of 414 - 2302 MPa to permit AIM of 0.2 - 1 mm thereby reducing DH.

Biomechanical Analysis Using FEA and Experiments of Metal Plate and Bone Strut Repair of a Femur Midshaft Segmental Defect

This investigation assessed the biomechanical performance of the metal plate and bone strut technique for fixing recalcitrant nonunions of femur midshaft segmental defects, which has not been systematically done before. A finite element (FE) model was developed and then validated by experiments with the femur in 15 deg of adduction at a subclinical hip force of 1 kN. Then, FE analysis was done with the femur in 15 deg of adduction at a hip force of 3 kN representing about 4 x body weight for a 75 kg person to examine clinically relevant cases, such as an intact femur plus 8 different combinations of a lateral metal plate of fixed length, a medial bone strut of varying length, and varying numbers and locations of screws to secure the plate and strut around a midshaft defect. Using the traditional “high stiffness” femur-implant construct criterion, the repair technique using both a lateral plate and a medial strut fixed with the maximum possible number of screws would be the most desirable since it had the highest stiffness (1948 N/mm); moreover, this produced a peak femur cortical Von Mises stress (92 MPa) which was below the ultimate tensile strength of cortical bone. Conversely, using the more modern “low stiffness” femur-implant construct criterion, the repair technique using only a lateral plate but no medial strut provided the lowest stiffness (606 N/mm), which could potentially permit more in-line interfragmentary motion (i.e., perpendicular to the fracture gap, but in the direction of the femur shaft long axis) to enhance callus formation for secondary-type fracture healing; however, this also generated a peak femur cortical Von Mises stress (171 MPa) which was above the ultimate tensile strength of cortical bone.

Biomechanical analysis of the cephalomedullary nail versus the trochanteric stabilizing plate for unstable intertrochanteric femur fractures

Unstable intertrochanteric fractures are commonly treated with a cephalomedullary nail due to high failure rates with a sliding hip screw. The Omega3 Trochanteric Stabilizing Plate is a relatively new device that functions like a modified sliding hip screw with a proximal extension; however, its mechanical properties have not been evaluated. This study biomechanically compared a cephalomedullary nail, that is, Gamma3 Nail against the Omega3 plate. Unstable intertrochanteric fractures were created in 24 artificial femurs. Experimental groups were as follows: Nail (i.e. Gamma3 Nail) (n = 8), Plate A (i.e. Omega3 plate with four distal non-locking screws and no proximal locking screws) (n = 8), Plate B (i.e. Plate A plus five proximal locking screws) (n = 8), Plate C (i.e. Omega3 plate with three distal locking screws and no proximal locking screws) (n = 8), and Plate D (i.e. Plate C plus five proximal locking screws) (n = 8). All specimens were stiffness tested, while the Nail and Plate D groups were also strength tested. For lateral bending, Plate B was less stiff than the Nail (p = 0.001) and Plate A (p = 0.009). For torsion, Plate A was less stiff than Plate D (p = 0.020). For axial compression, the Nail was less stiff than Plate A (p = 0.036) and Plate B (p = 0.008). Axial strength for the Nail (5014 ± 308 N) was 66% higher than the Plate D construct (2940 ± 411 N) (p < 0.001). All Nails failed by partial or complete cutout through the femoral head and neck, but Plate D failed by varus collapse and deformation of the lag screw. When the cephalomedullary nail is clinically contra-indicated, this study supports the use of the Omega3 plate, since it had similar stiffness in three test modes to the Gamma3 Nail, but had lower strength. Stability of Omega3 plate constructs was not improved with locked fixation proximally or distally.

Periprosthetic fractures: concepts of biomechanical in vitro investigations

PURPOSE

Experimental in vitro studies investigating periprosthetic fractures after joint replacement are used increasingly. The purpose of this review was to deliver a condensed survey of studies in order to provide researchers with an overview of relevant scientific results and their clinical relevance.

METHODS

A literature search was conducted to obtain all available papers dealing with periprosthetic fractures, with particular attention being paid to articles with an experimental research design. Study goals, scientific methods and results, their interpretation and clinical relevance were assessed and compared. The main focus was on comparability with clinical fracture patterns and physiological joint loads.

RESULTS

Excluding duplicates, 24 studies with regard to artificial hip, knee and shoulder joints were found dating back to August 2000. Almost all studies were performed quasi-statically and without consideration of muscle forces and thus reflect selected loading conditions and no dynamic situation during activities of daily living (ADL). Various experimental protocols were used, differing in the choice of experimental material, implant and fixation system and load application.

CONCLUSIONS

In vitro studies regarding periprosthetic fracture research allow controlling for disturbances, such as clinically occurring risk factors like reduced bone mineral density (BMD) or greater patient age. Notwithstanding, due to methodological differences, comparisons between studies were possible to a limited degree only. For this reason, and because of quasi-static loading typically applied, results can only be partially applied to clinical practice.

Biomechanical analysis of a new carbon fiber/flax/epoxy bone fracture plate shows less stress shielding compared to a standard clinical metal plate

Femur fracture at the tip of a total hip replacement (THR), commonly known as Vancouver B1 fracture, is mainly treated using rigid metallic bone plates which may result in "stress shielding" leading to bone resorption and implant loosening. To minimize stress shielding, a new carbon fiber (CF)/Flax/Epoxy composite plate has been developed and biomechanically compared to a standard clinical metal plate. For fatigue tests, experiments were done using six artificial femurs cyclically loaded through the femoral head in axial compression for four stages: Stage 1 (intact), stage 2 (after THR insertion), stage 3 (after plate fixation of a simulated Vancouver B1 femoral midshaft fracture gap), and stage 4 (after fracture gap healing). For fracture fixation, one group was fitted with the new CF/Flax/Epoxy plate (n = 3), whereas another group was repaired with a standard clinical metal plate (Zimmer, Warsaw, IN) (n = 3). In addition to axial stiffness measurements, infrared thermography technique was used to capture the femur and plate surface stresses during the testing. Moreover, finite element analysis (FEA) was performed to evaluate the composite plate's axial stiffness and surface stress field. Experimental results showed that the CF/Flax/Epoxy plated femur had comparable axial stiffness (fractured = 645 ± 67 N/mm; healed = 1731 ± 109 N/mm) to the metal-plated femur (fractured = 658 ± 69 N/mm; healed = 1751 ± 39 N/mm) (p = 1.00). However, the bone beneath the CF/Flax/Epoxy plate was the only area that had a significantly higher average surface stress (fractured = 2.10 ± 0.66 MPa; healed = 1.89 ± 0.39 MPa) compared to bone beneath the metal plate (fractured = 1.18 ± 0.93 MPa; healed = 0.71 ± 0.24 MPa) (p < 0.05). FEA bone surface stresses yielded peak of 13 MPa at distal epiphysis (stage 1), 16 MPa at distal epiphysis (stage 2), 85 MPa for composite and 129 MPa for metal-plated femurs at the vicinity of nearest screw just proximal to fracture (stage 3), 21 MPa for composite and 24 MPa for metal-plated femurs at the vicinity of screw farthest away distally from fracture (stage 4). These results confirm that the new CF/Flax/Epoxy material could be a potential candidate for bone fracture plate applications as it can simultaneously provide similar mechanical stiffness and lower stress shielding (i.e., higher bone stress) compared to a standard clinical metal bone plate.

Biomechanical measurements of stopping and stripping torques during screw insertion in five types of human and artificial humeri

During orthopedic surgery, screws are inserted by “subjective feel” in humeri for fracture fixation, that is, stopping torque, while trying to prevent accidental over-tightening that causes screw–bone interface failure, that is, stripping torque. However, no studies exist on stopping torque, stripping torque, or stopping/stripping torque ratio in human or artificial humeri. This study evaluated five types of humeri, namely, human fresh-frozen (n = 19), human embalmed (n = 18), human dried (n = 15), artificial “normal” (n = 13), and artificial “osteoporotic” (n = 13). An orthopedic surgeon used a torque screwdriver to insert 3.5-mm-diameter cortical screws into humeral shafts and 6.5-mm-diameter cancellous screws into humeral heads by “subjective feel” to obtain stopping and stripping torques. The five outcome measures were raw and normalized stopping torque, raw and normalized stripping torque, and stopping/stripping torque ratio. Normalization was done as raw torque/screw–bone interface area. For “gold standard” fresh-frozen humeri, cortical screw tests yielded averages of 1312 N mm (raw stopping torque), 30.4 N/mm (normalized stopping torque), 1721 N mm (raw stripping torque), 39.0 N/mm (normalized stripping torque), and 82% (stopping/stripping torque ratio). Similarly, fresh-frozen humeri gave cancellous screw average results of 307 N mm (raw stopping torque), 0.9 N/mm (normalized stopping torque), 392 N mm (raw stripping torque), 1.2 N/mm (normalized stripping torque), and 79% (stopping/stripping torque ratio). Of the five cortical screw parameters for fresh-frozen humeri versus other groups, statistical equivalence (p ≥ 0.05) occurred in four cases (embalmed), three cases (dried), four cases (artificial “normal”), and four cases (artificial “osteoporotic”). Of the five cancellous screw parameters for fresh-frozen humeri versus other groups, statistical equivalence (p ≥ 0.05) occurred in five cases (embalmed), one case (dried), one case (artificial “normal”), and zero cases (artificial “osteoporotic”). Stopping/stripping torque ratios were relatively constant for all groups at 77%–88% (cortical screws) and 79%–92% (cancellous screws).

The biomechanical effect of proximal tumor defect location on femur pathological fractures

OBJECTIVES

The femur is the most common long bone affected by cancerous metastasis. Femoral tumor defects are known to induce pain and functional impairment in patients. Although prior studies exist evaluating the clinical and biomechanical effect of tumor defect size, no biomechanical studies have experimentally examined the risk of pathological fracture with respect to the anterior, posterior, medial, and lateral surfaces on which a proximal tumor defect is located on the femur.

METHODS

Circular tumor-like defects of 40-mm diameter were created proximally in the subtrochanteric region on the Anterior (n = 5), Posterior (n = 5), Medial (n = 5), and Lateral (n = 5) sides of 20 synthetic femurs. Intact femurs served as a control group (n = 4). Femurs were tested for lateral, "offset" torsional, and axial stiffness, as well as axial strength.

RESULTS

Lateral stiffnesses (range, 121-162 N/mm) yielded no differences between groups (P = 0.069). "Offset" torsional stiffnesses (range, 135-188 N/mm) demonstrated that the Medial group was less stiff than the Intact, Anterior, and Lateral groups (P ≤ 0.012). Axial stiffnesses (range, 1057-1993 N/mm) showed that the Medial group was less stiff than the Intact group (P = 0.006). Axial strengths (range, 3250-6590 N) for the Medial group were lower than Anterior (P = 0.001) and Posterior (P = 0.001) specimens, whereas the Lateral group had a lower strength than Anterior specimens (P = 0.019). No other statistical differences were noted. Axial failure of Medial and Lateral specimens involved the tumor-like defect in 100% of cases, whereas 100% of Intact femurs and 80% of Anterior and Posterior femur groups failed only through the neck.

CONCLUSIONS

In 2 of 3 test modes, the Medial tumor-like defect group resulted in statistically lower stiffness values compared with Intact femurs and had lower strength than Anterior and Posterior groups in axial failure.

The biomechanical effect of notch size, notch location, and femur orientation on hip resurfacing failure

For hip resurfacing, this is the first biomechanical study to assess anterior and posterior femoral neck notching and femur flexion and extension. Forty-seven artificial femurs were implanted with the Birmingham hip resurfacing (BHR) using a range of notch sizes (0, 2, and 5 mm), notch locations (superior, anterior, and posterior), and femur orientations (neutral stance, flexion, and extension). Implant preparation was done using imageless computer navigation, and mechanical tests measured stiffness and strength. For notch size and location, in neutral stance the unnotched group had 1.9 times greater strength than the 5-mm superior notch group (4539 N versus 2423 N, p=0.047), and the 5-mm anterior notch group had 1.6 times greater strength than the 5-mm superior notch group, yielding a borderline statistical difference (3988 N versus 2423 N, p = 0.056). For femur orientation, in the presence of a 5-mm anterior notch, femurs in neutral stance had 2.2 times greater stiffness than femurs in 25° flexion (1542 N/mm versus 696 N/mm, p = 0.000). Similarly, in the presence of a 5-mm posterior notch, femurs in neutral stance had 2.8 times greater stiffness than femurs in 25° extension (1637 N/mm versus 575 N/mm, p = 0.000). No other statistical differences were noted. All femurs failed through the neck. The results have implications for BHR surgical techniques and recommended patient activities.

A biomechanical comparison of four different cementless press-fit stems used in revision surgery for total knee replacements

Few biomechanical studies exist on femoral cementless press-fit stems for revision total knee replacement (TKR) surgeries. The aim of this study was to compare the mechanical quality of the femur-stem interface for a series of commercially available press-fit stems, because this interface may be a 'weak link' which could fail earlier than the femur-TKR bond itself. Also, the femur-stem interface may become particularly critical if distal femur bone degeneration, which may necessitate or follow revision TKR, ever weakens the femur-TKR bond itself. The authors implanted five synthetic femurs each with a Sigma Short Stem (SSS), Sigma Long Stem (SLS), Genesis II Short Stem (GSS), or Genesis II Long Stem (GLS). Axial stiffness, lateral stiffness, 'offset load' torsional stiffness, and 'offset load' torsional strength were measured with a mechanical testing system using displacement control. Axial (range = 1047-1461 N/mm, p = 0.106), lateral (range = 415-462 N/mm, p = 0.297), and torsional (range = 115-139 N/mm, p > 0.055) stiffnesses were not different between groups. The SSS had higher torsional strength (863 N) than the other stems (range = 167-197 N, p < 0.001). Torsional failure occurred by femoral 'spin' around the stem's long axis. There was poor linear correlation between the femur-stem interface area versus axial stiffness (R = 0.38) and torsional stiffness (R = 0.38), and there was a moderate linear correlation versus torsional strength (R = 0.55). Yet, there was a high inverse linear correlation between interfacial surface area versus lateral stiffness (R = 0.79), although this did not result in a statistical difference between stem groups (p = 0.297). These press-fit stems provide equivalent stability, except that the SSS has greater torsional strength.

Predictors of femoral neck fracture following hip resurfacing: a cadaveric study

We aimed to establish if radiological parameters, dual energy x-ray absorptiometry (DEXA) and quantitative CT (qCT) could predict the risk of sustaining a femoral neck fracture following hip resurfacing. Twenty-one unilateral fresh frozen femurs were used. Each femur had a plain digital anteroposterior radiograph, DEXA scan and qCT scan. Femurs were then prepared for a Birmingham Hip Resurfacing femoral component and loaded to failure. Results demonstrated that gender and qCT measurements showed strong correlation with failure load. QCT could be used as an individual measure to predict risk of post-operative femoral neck fracture. However, when qCT is unavailable; gender, pre-operative DEXA scan and Neck Width measurements can be used together to assess risk of post-operative femoral neck fracture in patients due to undergo hip resurfacing.

Biomechanical Measurements of Surgical Drilling Force and Torque in Human Versus Artificial Femurs

Few experimental studies have examined surgical drilling in human bone, and no studies have inquired into this aspect for a popular commercially-available artificial bone used in biomechanical studies. Sixteen fresh-frozen human femurs and five artificial femurs were obtained. Cortical specimens were mounted into a clamping system equipped with a thrust force and torque transducer. Using a CNC machine, unicortical holes were drilled in each specimen at 1000 rpm, 1250 rpm, and 1500 rpm with a 3.2 mm diameter surgical drill bit. Feed rate was 120 mm/min. Statistical significance was set at p < 0.05. Force at increasing spindle speed (1000 rpm, 1250 rpm, and 1500 rpm), respectively, showed a range for human femurs (198.4 ± 14.2 N, 180.6 ± 14.0 N, and 176.3 ± 11.2 N) and artificial femurs (87.2 ± 19.3 N, 82.2 ± 11.2 N, and 75.7 ± 8.8 N). For human femurs, force at 1000 rpm was greater than at other speeds (p ≤ 0.018). For artificial femurs, there was no speed effect on force (p ≥ 0.991). Torque at increasing spindle speed (1000 rpm, 1250 rpm, and 1500 rpm), respectively, showed a range for human femurs (186.3 ± 16.9 N·mm, 157.8 ± 16.1 N·mm, and 140.2 ± 16.4 N·mm) and artificial femurs (67.2 ± 8.4 N·mm, 61.0 ± 2.9 N·mm, and 53.3 ± 2.9 N·mm). For human femurs, torque at 1000 rpm was greater than at other speeds (p < 0.001). For artificial femurs, there was no difference in torque for 1000 rpm versus higher speeds (p ≥ 0.228), and there was only a borderline difference between the higher speeds (p = 0.046). Concerning human versus artificial femurs, their behavior was different at every speed (force, p ≤ 0.001; torque, p < 0.001). For human specimens at 1500 rpm, force and torque were linearly correlated with standardized bone mineral density (sBMD) and the T-score used to clinically categorize bone quality (R ≥ 0.56), but there was poor correlation with age at all speeds (R ≤ 0.37). These artificial bones fail to replicate force and torque in human cortical bone during surgical drilling. To date, this is the largest series of human long bones biomechanically tested for surgical drilling.

The biomechanical effect of torsion on humeral shaft repair techniques for completed pathological fractures

In the presence of a tumor defect, completed humeral shaft fractures continue to be a major surgical challenge since there is no "gold standard" treatment. This is due, in part, to the fact that only one prior biomechanical study exists on the matter, but which only compared 2 repair methods. The current authors measured the humeral torsional performance of 5 fixation constructs for completed pathological fractures. In 40 artificial humeri, a 2-cm hemi-cylindrical cortical defect with a transverse fracture was created in the lateral cortex. Specimens were divided into 5 different constructs and tested in torsion. Construct A was a broad 10-hole 4.5-mm dynamic compression plate (DCP). Construct B was the same as A except that the screw holes and the tumor defect were filled with bone cement and the screws were inserted into soft cement. Construct C was the same as A except that the canal and tumor defect were filled with bone cement and the screws were inserted into dry cement. Construct D was a locked intramedullary nail inserted in the antegrade direction. Construct E was the same as D except that bone cement filled the defect. For torsional stiffness, construct C (4.45 ± 0.20 Nm/deg) was not different than B or E (p > 0.16), but was higher than A and D (p < 0.001). For failure torque, construct C achieved a higher failure torque (69.65 ± 5.35 Nm) than other groups (p < 0.001). For the failure angle, there were no differences between plating constructs A to C (p ≥ 0.11), except for B versus C (p < 0.05), or between nailing groups D versus E (p = 0.97), however, all plating groups had smaller failure angles than both nailing groups (p < 0.05). For failure energy, construct C (17.97 ± 3.59 J) had a higher value than other groups (p < 0.005), except for A (p = 0.057). Torsional failure always occurred in the bone in the classic "spiral" pattern. Construct C provided the highest torsional stability for a completed pathological humeral shaft fracture.

A preliminary biomechanical study of cyclic preconditioning effects on canine cadaveric whole femurs

Biomechanical preconditioning of biological specimens by cyclic loading is routinely done presumably to stabilize properties prior to the main phase of a study. However, no prior studies have actually measured these effects for whole bone of any kind. The aim of this study, therefore, was to quantify these effects for whole bones. Fourteen matched pairs of fresh-frozen intact cadaveric canine femurs were sinusoidally loaded in 4-point bending from 50 N to 300 N at 1 Hz for 25 cycles. All femurs were tested in both anteroposterior (AP) and mediolateral (ML) bending planes. Bending stiffness (i.e., slope of the force-vs-displacement curve) and linearity R(2) (i.e., coefficient of determination) of each loading cycle were measured and compared statistically to determine the effect of limb side, cycle number, and bending plane. Stiffnesses rose from 809.7 to 867.7 N/mm (AP, left), 847.3 to 915.6 N/mm (AP, right), 829.2 to 892.5 N/mm (AP, combined), 538.7 to 580.4 N/mm (ML, left), 568.9 to 613.8 N/mm (ML, right), and 553.8 to 597.1 N/mm (ML, combined). Linearity R(2) rose from 0.96 to 0.99 (AP, left), 0.97 to 0.99 (AP, right), 0.96 to 0.99 (AP, combined), 0.95 to 0.98 (ML, left), 0.94 to 0.98 (ML, right), and 0.95 to 0.98 (ML, combined). Stiffness and linearity R(2) versus cycle number were well-described by exponential curves whose values leveled off, respectively, starting at 12 and 5 cycles. For stiffness, there were no statistical differences for left versus right femurs (p = 0.166), but there were effects due to cycle number (p < 0.0001) and AP versus ML bending plane (p < 0.0001). Similarly, for linearity, no statistical differences were noted due to limb side (p = 0.533), but there were effects due to cycle number (p < 0.0001) and AP versus ML bending plane (p = 0.006). A minimum of 12 preconditioning cycles was needed to fully stabilize both the stiffness and linearity of the canine femurs. This is the first study to measure the effects of mechanical preconditioning on whole bones, having some practical implications on research practices.

The biomechanics of three different fracture fixation implants for distal femur repair in the presence of a tumor-like defect

The femur is the most common long bone involved in metastatic disease. There is consensus about treating diaphyseal and epiphyseal metastatic lesions. However, the choice of device for optimal fixation for distal femur metaphyseal metastatic lesion remains unclear. This study compared the mechanical stiffness and strength of three different fixation methods. In 15 synthetic femurs, a spherical tumor-like defect was created in the lateral metaphyseal region, occupying 50% of the circumference of the bone. The defect was filled with bone cement and fixed with one of three methods: Group 1 (retrograde nail), Group 2 (lateral locking plate), and Group 3 (lateral nonlocking periarticular plate). Constructs were tested for mechanical stiffness and strength. There were no differences between groups for axial stiffness (Group 1, 1280 ± 112 N/mm; Group 2, 1422 ± 117 N/mm; and Group 3, 1403 ± 122 N/mm; p = 0.157) and offset torsional strength (Group 1, 1696 ± 628 N; Group 2, 1771 ± 290 N; and Group 3, 1599 ± 253 N; p = 0.816). In the coronal plane, Group 2 (296 ± 17 N/mm) had a higher stiffness than Group 1 (263 ± 17 N/mm; p = 0.018). In the sagittal plane, Group 1 (315 ± 9 N/mm) had a higher stiffness than Group 3 (285 ± 19 N/mm; p = 0.028). For offset torsional stiffness, Group 1 (256 ± 23 N/mm) had a higher value than Group 3 (218 ± 16 N/mm; p = 0.038). Group 1 had equivalent performance to both plating groups in two test modes, and it was superior to Group 3 in two other test modes. Since a retrograde nail (i.e. Group 1) would require less soft-tissue stripping in a clinical context, it may be the optimal choice for tumor-like defects in the distal femur.

Biomechanical measurements of cortical screw stripping torque in human versus artificial femurs

Femur fracture plates are applied using cortical bone screws. Surgeons do this manually by subjective ‘feel’ without monitoring torque. Few studies have quantified stripping torque in human bone. No studies have measured stripping torque in the artificial bones from Sawbones (Vashon, WA, USA) that are frequently used in biomechanical studies. The present aim was to measure stripping torque of cortical screws in human versus artificial femurs. Sixteen fresh-frozen human femurs and eight artificial femurs were used. Using a digital torque screwdriver, each femur had a 3.5-mm diameter unicortical screw manually inserted into the anterior midshaft until failure of the screw–bone interface. Results were normalized by cortical thickness and the screw–bone interfacial area. There were no statistical differences in human versus artificial data, respectively, for stripping torque (1741 ± 442 N.mm, 2012 ± 176 N.mm, p = 0.11), stripping torque/thickness (313 ± 59 N, 305 ± 30 N, p = 0.74), and stripping torque/area (28.5 ± 5.3 N/mm, 27.8 ± 2.8 N/mm, p = 0.74). Artificial unicortical thickness (6.6 ± 0.3 mm) was greater than human thickness (5.6 ± 1.1 mm) ( p = 0.02). For human specimens, there was a moderate linear correlation of absolute and normalized stripping torque versus standardized bone mineral density (R ≥ 0.32) and clinical T-score (R = 0.29), but not with age (R ≤ 0.29). Surgeons should be aware of the stripping torque limits for human femurs and potentially take steps to monitor these values during surgery. The artificial femurs being increasingly used in research accurately replicate human cortical properties during screw insertion. To date, this is the first series of human femurs evaluated for cortical screw stripping.

Biomechanical stress maps of an artificial femur obtained using a new infrared thermography technique validated by strain gages

Femurs are the heaviest, longest, and strongest long bones in the human body and are routinely subjected to cyclic forces. Strain gages are commonly employed to experimentally validate finite element models of the femur in order to generate 3D stresses, yet there is little information on a relatively new infrared (IR) thermography technique now available for biomechanics applications. In this study, IR thermography validated with strain gages was used to measure the principal stresses in the artificial femur model from Sawbones (Vashon, WA, USA) increasingly being used for biomechanical research. The femur was instrumented with rosette strain gages and mechanically tested using average axial cyclic forces of 1500 N, 1800 N, and 2100 N, representing 3 times body weight for a 50 kg, 60 kg, and 70 kg person. The femur was oriented at 7° of adduction to simulate the single-legged stance phase of walking. Stress maps were also obtained using an IR thermography camera. Results showed good agreement of IR thermography vs. strain gage data with a correlation of R(2)=0.99 and a slope=1.08 for the straight line of best fit. IR thermography detected the highest principal stresses on the superior-posterior side of the neck, which yielded compressive values of -91.2 MPa (at 1500 N), -96.0 MPa (at 1800 N), and -103.5 MPa (at 2100 N). There was excellent correlation between IR thermography principal stress vs. axial cyclic force at 6 locations on the femur on the lateral (R(2)=0.89-0.99), anterior (R(2)=0.87-0.99), and posterior (R(2)=0.81-0.99) sides. This study shows IR thermography's potential for future biomechanical applications.

Biomechanical measurements of axial crush injury to the distal condyles of human and synthetic femurs

Few studies have evaluated the ‘bulk’ mechanical properties of human longbones and even fewer have compared human tissue to the synthetic longbones increasingly being used by researchers. Distal femur fractures, for example, comprise about 6% of all femur fractures, but the mechanical properties of the distal condyles of intact human and synthetic femurs have not been well quantified in the literature. To this end, the distal portions of a series of 16 human fresh-frozen femurs and six synthetic femurs were prepared identically for mechanical testing. Using a flat metal plate, an axial ‘crush’ force was applied in-line with the long axis of the femurs. The two femur groups were statistically compared and values correlated to age, size, and bone quality. Results yielded the following: crush stiffness (human, 1545 ± 728 N/mm; synthetic, 3063 ± 1243 N/mm; p = 0.002); crush strength (human, 10.3 ± 3.1 kN; synthetic, 12.9 ± 1.7 kN; p = 0.074); crush displacement (human, 6.1 ± 1.8 mm; synthetic, 2.8 ± 0.3 mm; p = 0.000); and crush energy (human, 34.8 ± 15.9 J; synthetic, 18.1 ± 5.7 J; p = 0.023). For the human femurs, there were poor correlations between mechanical properties versus age, size, and bone quality (R2 ≤ 0.18), with the exception of crush strength versus bone mineral density (R2 = 0.33) and T-score (R2 = 0.25). Human femurs failed mostly by condyle ‘roll back’ buckling (15 of 16 cases) and/or unicondylar or bicondylar fracture (7 of 16 cases), while synthetic femurs all failed by wedging apart of the condyles resulting in either fully or partially displaced condylar fractures (6 of 6 cases). These findings have practical implications on the use of a flat plate load applicator to reproduce real-life clinical failure modes of human femurs and the appropriate use of synthetic femurs. To the authors’ knowledge, this is the first study to have done such an assessment on human and synthetic femurs.