Pixel-Based DXA-Derived Structural Properties Strongly Correlate with pQCT Measures at the One-Third Distal Femur Site.

While bone mineral density has been traditionally used to quantify fracture risk for individuals with spinal cord injuries, recent studies are including engineering measurements such as section modulus and cross sectional moment of inertia. These are almost exclusively calculated by peripheral QCT scanners which, unlike DXA scanners, are rarely found in clinical settings. Using fifty-four fresh frozen femora, we developed and validated a pixel-by-pixel method to calculate engineering properties at the distal femur using a Hologic QDR-1000 W DXA scanner and compared them against similar parameters measured using a Stratec XCT-3000 peripheral QCT scanner. We found excellent agreement between standard DXA and pixel-by-pixel measured BMD (r 2 = 0.996). Cross-sectional moment of inertia about the anteroposterior axis measured using DXA and pQCT correlated very strongly (r 2 = 0.99). Cross-sectional moment of inertia about the anteroposterior axis measured using DXA also correlated strongly with pQCT measured bone strength index (r 2 = 0.99). These correlations indicate that DXA scans can measure equivalent pQCT parameters, and some existing DXA scans can be reprocessed with pixel-by-pixel techniques. Ultimately, these engineering parameters may help better quantify fracture-risk in fracture-prone populations such as those with spinal cord injuries.

Comparison of three methods of calculating strain in the mouse ulna in exogenous loading studies.

Axial compression of mouse limbs is commonly used to induce bone formation in a controlled, non-invasive manner. Determination of peak strains caused by loading is central to interpreting results. Load-strain calibration is typically performed using uniaxial strain gauges attached to the diaphyseal, periosteal surface of a small number of sacrificed animals. Strain is measured as the limb is loaded to a range of physiological loads known to be anabolic to bone. The load-strain relationship determined by this subgroup is then extrapolated to a larger group of experimental mice. This method of strain calculation requires the challenging process of strain gauging very small bones which is subject to variability in placement of the strain gauge. We previously developed a method to estimate animal-specific periosteal strain during axial ulnar loading using an image-based computational approach that does not require strain gauges. The purpose of this study was to compare the relationship between load-induced bone formation rates and periosteal strain at ulnar midshaft using three different methods to estimate strain: (A) Nominal strain values based solely on load-strain calibration; (B) Strains calculated from load-strain calibration, but scaled for differences in mid-shaft cross-sectional geometry among animals; and (C) An alternative image-based computational method for calculating strains based on beam theory and animal-specific bone geometry. Our results show that the alternative method (C) provides comparable correlation between strain and bone formation rates in the mouse ulna relative to the strain gauge-dependent methods (A and B), while avoiding the need to use strain gauges.

Component alignment during total knee arthroplasty with use of standard or custom instrumentation: a randomized clinical trial using computed tomography for postoperative alignment measurement.

BACKGROUND: Patient-specific femoral and tibial cutting blocks produced with use of data from preoperative computed tomography (CT) or magnetic resonance imaging (MRI) scans have been employed recently to optimize component alignment in total knee arthroplasty. We report the results of a randomized controlled trial in which CT scans were used to compare postoperative component alignment between patients treated with custom instruments and those managed with traditional instruments. METHODS: The in-hospital data and early clinical outcomes, including Knee Society scores, were determined in a randomized clinical trial of forty-seven patients who had undergone a total of forty-eight primary total knee arthroplasties with patient-specific instruments (twenty-two knees) or standard instruments (twenty-six knees). Orientation of the implants was compared by using three-dimensional CT data. RESULTS: No significant differences were found between the study and control groups with respect to any clinical outcome after a minimum of six months of follow-up. The patient-specific tibial cutting block was abandoned in favor of a standard external alignment jig in seven of the twenty-two study knees because of possible malalignment. A detailed analysis of intent-to-treat and per-protocol groups of study and control knees did not show any significant improvement in component alignment, including femoral component rotation in the axial plane, in the patients treated with the custom instruments. The percentage of outliers--defined as less than -3° or more than 3° from the correct orientation of the tibial slope--was significantly higher in the group treated with use of patient-specific blocks than it was in the control group, in both the intent-to-treat (32% versus 8%, p = 0.032) and the per-protocol (47% versus 6%, p = 0.0008) analysis. CONCLUSIONS: There were no significant improvements in clinical outcomes or knee component alignment in patients treated with patient-specific cutting blocks as compared with those treated with standard instruments. The group treated with patient-specific cutting blocks had a significantly higher prevalence of malalignment in terms of tibial component slope than the knees treated with standard instruments.

Geometric mouse variation: implications to the axial ulnar loading protocol and animal specific calibration.

Large variations in axial ulnar load strain calibration results suggest that animal-specific calibrations may be necessary. However, the optimal set of geometric measures for performing an animal-specific calibration are not known, potentially as a result of confounding effects associated with experimentally introduced variation. The purpose of this study was to characterize the inherent variability of ulnar geometric measures known to influence periosteal midshaft strain during axial ulnar exogenous loading, and to further quantify the relationship between the variance of those geometric measures and periosteal strain during axial loading. Thirty-nine right mouse forelimbs were scanned with microCT. Seven geometric measures that influence periosteal strain resulting from a combined axial and bending loading were computed and used to estimate animal-specific strains on the periosteal midshaft. Animal specific strains were estimated using a theoretical model based on the generalized flexure formula. The predicted mean and standard deviation of the simulated midshaft strain gauge measurement resulting from the inter-animal geometric differences was -985 ± 148 µÎµ/N. The complete beam bending term associated with bending about the I(min) axis accounted for 89% of the variance and reduced the residual RMSE to 50.4 µÎµ. Eccentricity associated with the axial loading contributed a substantial portion of variation to the computed strain suggesting that calibration procedures to account for animal differences should incorporate that variable. The method developed in this study provides a relatively simple procedure for computing animal-specific strains using microCT scan data, without the need of a load/strain calibration study or computationally intensive finite element models.

Consistency among musculoskeletal models: caveat utilitor.

Musculoskeletal simulation software and model repositories have broadened the user base able to perform musculoskeletal analysis and have facilitated in the sharing of models. As the recognition of musculoskeletal modeling continues to grow as an engineering discipline, the consistency in results derived from different models and software is becoming more critical. The purpose of this study was to compare eight models from three software packages and evaluate differences in quadriceps moment arms, predicted muscle forces, and predicted tibiofemoral contact forces for an idealized knee-extension task spanning -125 to +10° of knee extension. Substantial variation among models was observed for the majority of aspects evaluated. Differences among models were influenced by knee angle, with better agreement of moment arms and tibiofemoral joint contact force occurring at low to moderate knee flexion angles. The results suggest a lack of consistency among models and that output differences are not simply an artifact of naturally occurring inter-individual differences. Although generic musculoskeletal models can easily be scaled to consistent limb lengths and use the same muscle recruitment algorithm, the results suggest those are not sufficient conditions to produce consistent muscle or joint contact forces, even for simplified models with no potential of co-contraction.

Step scaling and behaviour selection in a constrained set of manual material handling transfers.

Predictive biomechanical analysis of manual material handling (MMH) transfers is dependent on accurate prediction of foot locations relative to the task. Previous studies have classified common acyclic stepping patterns used during those transfer tasks, but the influence of walking distance prior to the transfer is not well understood. Twenty men and women performed transfers for a minimum of six different delivery distance conditions. The number of steps used by the participants ranged from two to seven. A theoretical framework for idealised step-scaling strategies is proposed and compared with the experimental data. The maximum observed increase in step length prior to delivery was 1.43 times the nominal step length calculated for each participant. The data suggest that although participants can scale their steps to facilitate the use of a single terminal stance at the transfer, the majority of participants chose to utilise a combination of stepping strategies if the preferred contralateral lead foot strategy could not be easily implemented. PRACTITIONER SUMMARY: Accurate foot placements are needed for predictive biomechanical analysis of MMH. A laboratory study investigated the influence of previous step positions on MMH. A flexible step-scaling strategy, in which step lengths and strategy were varied, suggests that analysis based on simulated movements should consider multiple lifting postures.

Replicating a Colles fracture in an excised radius: revisiting testing protocols.

A distal radius fracture in middle-age and older adults is often considered a sentinel indicator of osteoporosis. Mechanical testing of cadaveric specimens is often used to quantify bone strength and develop insight for relating in-vivo measures to fracture force. Mechanical testing protocols using an intact forearm have been successful at replicating a Colles fracture, however, excised isolated radius protocols based on the intact forearm testing protocol have not been as successful. One protocol originally designed to replicate the physiological condition of a fall on an outstretched hand was reproduced in our laboratory, yet surprisingly the produced distal radius fracture patterns were not consistent among specimens nor was dorsal angulation of the distal fragment that is characteristic of a Colles fracture observed. The purpose of this study was to perform a mechanics-based analysis of the excised radius loading protocol in order to quantify the imposed and internal forces on the radius. An idealized beam model of the excised radius revealed that in the area of the distal radius where Colles fractures occur, 99.99% of the maximum strain on the bone outer surface was the result of pure compressive loading. This loading condition is in direct contrast to the accepted mechanics of a Colles fracture, which is characterized as a metaphyseal bending fracture with the volar cortex failing due to tensile stresses and the dorsal cortex exhibiting compression and comminution. The results suggest that additional research, particularly related to overcoming the difficulties of reliably supporting and applying a force to the distal end of the radius, is necessary for clinical fracture patterns to be reliably reproduced with an excised radius mechanical testing protocol.

Deriving tissue density and elastic modulus from microCT bone scans.

Tissue level density and elastic modulus are intrinsic properties that can be used to quantify bone material and analyses incorporating those quantities have been used to evaluate bone on a macroscopic scale. Micro-computed tomography (microCT) technology has been used to construct tissue level finite element models to simulate macroscopic fracture strength, however, a single method for assigning voxel-specific tissue density and elastic modulus based on those data has not been universally accepted. One method prevalent in the literature utilizes an empirical relationship that derives tissue stiffness as a function of bone calcium content weight fraction. To derive calcium content weight fraction from microCT scans, a measure of tissue density is required and a constant value is traditionally used. However, experimental data suggest a non-trivial amount of tissue heterogeneity suggesting a constant tissue density may not be appropriate. A theoretical derivation for determining the relationship between voxel-specific tissue density and microCT scan data (i.e., microCT derived tissue mineral density (TMD), mgHA/cm(3)) and bone constituent properties is proposed. Constant model parameters used in the derivation include the density of water, ash, and organics (i.e., bone constituents) and the volume fraction of the organics constituent. The effect of incorporating the theoretically derived tissue density (instead of a constant value) in determining voxel-specific elastic modulus resulted in a maximum observed increase of 12GPa (5.9GPa versus 17.9GPa, for the constant value and derived tissue density formulations, respectively) for a measured TMD of 1.02gHA/cm(3). Average and bounding quantities for the four constant model parameters were defined from the literature and the influence of those values on the derived tissue density and elastic modulus relationships were also evaluated. The theoretical relationships of tissue density and elastic modulus, with the average constant model parameters applied, were consistent with previously published empirical relationships derived from experimental data. Tissue density as a function of microCT TMD was formulated as a linear relationship and the density of water and ash was shown to solely influence the proportionality (i.e., slope) between those values. The density of water and organics (i.e., collagen) and the volume fraction of the organics constituent were shown to influence the constant offset (intercept) between tissue density and TMD with no influence from ash density. Incorporating tissue density heterogeneity into the derivation of elastic modulus resulted in a significant increase in predicted modulus (for microCT TMD ranges observed for healthy tissue) as compared to when a constant tissue density was used. The presented approach provides a novel method for deriving tissue-level bone material properties and quantifies the effect of assuming tissue homogeneity when calculating elastic modulus (when using a prevalent method in the literature) from microCT scan data.

The development of a model to predict the effects of worker and task factors on foot placements in manual material handling tasks.

Accurate prediction of foot placements in relation to hand locations during manual materials handling tasks is critical for prospective biomechanical analysis. To address this need, the effects of lifting task conditions and anthropometric variables on foot placements were studied in a laboratory experiment. In total, 20 men and women performed two-handed object transfers that required them to walk to a shelf, lift an object from the shelf at waist height and carry the object to a variety of locations. Five different changes in the direction of progression following the object pickup were used, ranging from 45° to 180° relative to the approach direction. Object weights of 1.0 kg, 4.5 kg, 13.6 kg were used. Whole-body motions were recorded using a 3-D optical retro-reflective marker-based camera system. A new parametric system for describing foot placements, the Quantitative Transition Classification System, was developed to facilitate the parameterisation of foot placement data. Foot placements chosen by the subjects during the transfer tasks appeared to facilitate a change in the whole-body direction of progression, in addition to aiding in performing the lift. Further analysis revealed that five different stepping behaviours accounted for 71% of the stepping patterns observed. More specifically, the most frequently observed behaviour revealed that the orientation of the lead foot during the actual lifting task was primarily affected by the amount of turn angle required after the lift (R(2) = 0.53). One surprising result was that the object mass (scaled by participant body mass) was not found to significantly affect any of the individual step placement parameters. Regression models were developed to predict the most prevalent step placements and are included in this paper to facilitate more accurate human motion simulations and ergonomics analyses of manual material lifting tasks. STATEMENT OF RELEVANCE: This study proposes a method for parameterising the steps (foot placements) associated with manual material handling tasks. The influence of task conditions and subject anthropometry on the foot placements of the most frequently observed stepping pattern during a laboratory study is discussed. For prospective postural analyses conducted using digital human models, accurate prediction of the foot placements is critical to realistic postural analyses and improved biomechanical job evaluations.

Foot motions in manual material handling transfer tasks: a taxonomy and data from an automotive assembly plant.

Ergonomic job analysis commonly applies static postural and biomechanical analysis tools to particular postures observed during manual material handling (MMH) tasks, usually focusing on the most extreme postures or those involving the highest loads. When these analyses are conducted prospectively using digital human models, accurate prediction of the foot placements is critical to realistic postural analyses. In automotive assembly jobs, workers frequently take several steps between task elements, for example, picking up a part at one location and moving to another location to place it on the vehicle. A detailed understanding of the influence of task type and task sequence on the stepping pattern is necessary to accurately predict the foot placements associated with MMH tasks. The current study examined the patterns of foot motions observed during automotive assembly tasks. Video data for 529 pickup and delivery tasks from 32 automotive assembly jobs were analysed. A minimum of five cycles was analysed for each task. The approach angle, departure angle, hand(s) used, manipulation height and patterns of footsteps were coded from the video. Object mass was identified from the job information sheet provided by the assembly plant. Three independent raters coded each video and demonstrated an intraclass correlation coefficient of 0.54 for identification of the configuration of the lower extremities during terminal stance. Based on an analysis of the distribution of stepping behaviours during object transitions (pickups or deliveries), a transition classification system (TRACS) was developed. TRACS uses a compact notation to quantify the sequence of steps associated with a MMH transition. Five TRACS behaviour groups accounted for over 90% of the transition stepping behaviours observed in the assembly plant. Approximately two-thirds (68.4%) of the object transfers observed were performed with only one foot in contact with the ground during the terminal posture. The results from this paper suggest that a predictive model for choosing a transition stepping behaviour, coupled with a model to scale the selected foot behaviours, is needed to facilitate accurate prospective ergonomic analyses. This study proposes a method for categorising the stepping patterns associated with MMH tasks. The influence of task type and task sequence on the stepping patterns observed during several automotive assembly tasks is discussed. For prospective postural analyses conducted using digital human models, accurate prediction of the foot placements is critical to realistic postural analyses.