Comparing Running-Specific and Traditional Prostheses During Running: Assessing Performance and Risk

Overview

The purpose of this research is to provide clinically, administratively, and field-relevant objective running outcomes by directly comparing running biomechanics of individuals with lower extremity amputation (ILEA) using RSPs (Running Specific Prostheses) and traditional prostheses. Within this purpose, the project has two specific aims: Specific Aim 1: To compare RSPs and traditional prostheses with respect to running ability and performance Specific Aim 2: To compare RSPs and traditional prostheses with respect to injury risks associated with running Hypothesis 1a: RSPs will outperform traditional prostheses at all velocities as measured by kinetic data (ground reaction forces, joint powers, joint and limb work) and 50m dash time. Hypothesis 1b: ILEA intact limbs and able-bodied control limbs will outperform residual limbs with RSPs and traditional prostheses at all velocities as measured by kinetic data. Hypothesis 2: Running with RSPs will show reduced acute and chronic injury risks compared to traditional prostheses at all velocities as measured by loading rates, EMG amplitudes, lumbopelvic kinematics, and modeled joint loads.

A. BACKGROUIND Current knowledge and understanding of individuals with lower extremity amputations (ILEA) when running are limited with respect to biomechanical performance and injury risks. ILEA are able to run with both running-specific prostheses (RSPs) and traditional prostheses; however, direct comparisons of subjects running with each of these prosthetic designs do not exist. The reported literature examining one design or the other often do not have participants running at the same velocities. Additionally, when running velocities are similar between studies, the same variables are rarely investigated. This makes comparisons between RSPs and traditional prostheses exceedingly difficult, and drawing conclusions on both performance and injury risk is virtually impossible. Furthermore, no ILEA running studies to date have investigated muscle activities, nor have running simulations of musculoskeletal models been generated. These major gaps in research substantially limit the understanding of both performance and injury risk of ILEA running with different prosthetic designs. Gaining this knowledge will directly inform clinicians and administrators within the Department of Defense and Veterans Administration systems on prosthesis prescription for running at a range of speeds as well as for return to duty scenarios. Therefore, the proposed study will utilize motion capture, muscle activity, and musculoskeletal modeling techniques to directly compare performance and injury risks of ILEA running with both RSPs and traditional prostheses across a range of speeds. The investigators will also capture an able-bodied control group for normative comparisons. In doing so, this project directly attends to the OPORA's goals and needs at multiple levels. B. RESEARCH STRATEGY: Subjects Twelve subjects with unilateral transtibial amputations will be recruited from the military, veteran, and civilian populations. Twelve able-bodied sex, age, height, and weight-matched subjects will serve as a control group to provide normative data for comparison. All subjects must provide their informed consent prior to beginning any portion of the study and will have the option to discontinue the study at any time without penalty. All subjects must complete the Physical Activity Readiness Questionnaire (PAR-Q) prior to beginning the experiment. If subjects fail this questionnaire (by responding "yes" to any of the 7 questions), they will not be allowed to participate in the study until they provide documentation from a physician clearing them to participate. Experimental Procedures A randomization computer program will determine the order of prosthetic foot conditions (RSP or traditional) for the ILEA subjects. All subjects (Controls and ILEA) will undergo 3D running analysis on an instrumented treadmill. A motion capture system will collect kinematic data (i.e. limb positions; joint angles, velocities, accelerations) at 200 Hz from reflective markers placed on specific body landmarks. On the amputated limb, the shank cluster will be placed laterally on the socket and a marker will be placed at the distal tip of the socket to define the long axis of the residual shank segment. For the RSP conditions, eight additional markers will be placed on the prosthesis keel. The marker on the most acute point of the prosthesis will define the prosthetic limb "ankle" joint. For the traditional prosthesis conditions, markers will be placed similarly to the intact foot. An instrumented treadmill will collect ground reaction force (GRF) data at 1000 Hz from force platforms imbedded in the treadmill. Kinematic and ground reaction force data will be combined and standard inverse dynamics techniques will be used to calculate joint kinetic data (i.e. forces, moments, powers, and work) using Visual3D (C-Motion, Germantown, MD) software. Muscle activation data will be collected at a sampling frequency of 2000 Hz using surface electromyography (EMG) from six bilateral hip and trunk muscles. Disposable pre-gelled silver-silver chloride (Ag-AgCl) electrodes will be placed over bilateral gluteus medius, gluteus maximus, rectus femoris, biceps femoris, lumbar erector spinae, internal oblique, and external oblique muscles. Maximal voluntary isometric contractions (MVIC) will be obtained for normalization purposes using manual resistance applied in standardized positions. All subjects (ILEA using RSPs, ILEA using traditional prosthesis, and Controls) will run at six prescribed speeds (2.5, 3.0, 3.5, 4.0, 5.0, 6.0 m/s) on a treadmill, presented in randomized order. ILEA subjects will complete all speed conditions wearing both their RSP and traditional prosthetic foot. As mentioned earlier, the prosthesis order will be randomized, but all speed conditions will be completed in one prosthesis prior to changing into the second prosthesis condition. This will save time and maintain a similar prosthetic fit between speed conditions. The study protocol requires that every subject runs at the same 6 specific speeds identified above. Each treadmill speed condition will take approximately 30 seconds, or the time needed for subjects to stabilize their gait at the speed and a collection of 10 consecutive strides to be collected. At least 2 minutes rest will be provided between each speed condition, and subjects can take more rest if needed. After completing the treadmill trials, subjects will complete 3x50m dashes in each foot condition to determine maximum speed. Again, at least 2 minutes rest will be provided between each effort, and subjects can take more rest if needed. The order of prosthesis use will also be randomized for this test, but all three dashes will be completed consecutively for each prosthesis condition. To evaluate performance while running, the primary outcome measures being evaluated will include peak joint powers; concentric, eccentric and total joint and limb work; average vertical and Anterior and Posterior (AP) GRFs; vertical and AP GRF impulses; and 50m dash time/speed. To evaluate injury risk while running, the primary outcome measures being evaluated will include average GRF magnitudes and loading rates, asymmetry in GRFs and joint moments, normalized EMG amplitudes, lumbopelvic kinematics, and peak and average joint contact forces over stance. Data for these measures will be collected for all subjects and then compared among the three groups (ILEA with RSPs, ILEA with Traditional Prosthesis, and Controls). No intervention will be offered to any subject and no follow-up to assess biomedical and/or health outcomes will be conducted. Electromyography Signal Analysis Raw EMG data will be demeaned, band-pass (cutoff frequency10-400 Hz) and band-stop (cutoff frequency 59-61 Hz) filtered to remove movement artifact and 60 Hz electrical noise using dual-pass (zero phase lag) Butterworth filters. EMG data will then be full wave rectified and low pass filtered to create linear envelopes. Peak values will be extracted from the MVIC trials to be used for normalization of the running trials to express EMG data as % Maximum Voluntary Contraction (MVC). Root mean square (RMS) of each muscle's activity will be calculated during stance and swing phases of running to investigate amplitude differences. Cross-correlation methodology will be utilized to investigate relative timing between trunk and hip musculature. Comparisons will be made to assess symmetry of muscle activation amplitude and timing between intact and prosthetic limb sides for each of the running speed and prosthesis conditions. Musculoskeletal Model Three musculoskeletal models will be developed in OpenSim. For the first model, the investigators will use a generic musculoskeletal model representing a non-amputee adult male, which has previously been used to assess muscle function during non-amputee running. The second and third models will be modified versions of the non-amputee model that represents a person using a prosthesis. The second model will include a traditional transtibial energy storage and return prosthesis, and the third model will include a running specific prosthesis. For both models, the investigators will remove the ankle muscles from the residual leg, including the medial and lateral gastrocnemius, soleus, tibialis posterior, flexor digitorum longus, tibialis anterior and extensor digitorum longus. Running Simulation Generation The models will be scaled to each individual subject using manual, uniform scale factors, computed from an inverse dynamics model based on kinematic marker positions and developed in Visual3D (C-Motion, Inc., Germantown, MD). Then, an inverse kinematics solution will be computed in Visual3D using a weighted least squares optimization algorithm. After the inverse kinematics solution has been computed, the investigators will perform a Residual Reduction Algorithm. Once minor changes have been made to the model and to the kinematic trajectories of the generalized coordinates, the investigators will implement a Computed Muscle Control algorithm. Once a solution for individual muscle excitations is found using the Computed Muscle Control algorithm, the investigators will compare it to the experimentally-collected electromyographic data. The investigators will generate running simulations at all running speeds for all participants. The results of these simulations will be compared between groups and across speeds, similar to the experimental measures that will be analyzed in this study. Joint Contact Force Analysis Once the running simulations have been developed, the three-dimensional contact load at the hip will be determined from the joint intersegmental force and compressive forces from the muscles crossing the joint, similar to the previous analyses during amputee walking. In addition, the muscle forces determined from the movement simulation will be used to help interpret the net contact loads. Material Properties and Anthropometrics Inertial properties of the prosthetic components and intact body segments will be estimated for use with the inverse dynamics approach. Subject masses will be measured using a force platform. Height and body weight of each subject will be measured, and anthropometric measurements from marker positions will be used to estimate the mass, center of mass, and moments of inertia of intact limb segments. Because ILEA subjects are missing one foot and part of their shank, an adjusted body mass (ABM) will be used as an input to anthropometric regression equations that account for the missing body segments. For subjects with amputation, the residual limb length and circumferences at the knee joint and distal end of the limb will be measured using a measuring tape. Residual limb inertial properties will then be estimated as a frustum of a right circular cone. Residual limb mass will be estimated from the calculated geometric volume assuming a uniform 1.10 g/cm3 tissue density. For the RSP conditions, the inertial properties of the RSP will be estimated from data published by the PI. Subsegments within the RSP keels will be defined via reflective marker placements and the inertial properties of each subsegment will be estimated by assuming each segment as a rigid trapezoidal cuboid. For the traditional prosthesis conditions, the inertial properties of the prosthesis will be estimated according to published literature. C. STATISTICAL PLAN and DATA ANALYSIS Statistical Analysis This research is designed to determine the effects of prosthesis type on a variety of performance and injury-risk related outcome variables across a range of running speeds. A three factor (Group x Limb x Speed) mixed model ANOVA will be used to identify statistical changes of the following dependent variables: peak hip, knee, and ankle joint powers and mechanical work; average ground reaction forces; loading rates; and EMG amplitudes. Group (Individual Lower Extremity Amputation Running-Specific Prosthesis (ILEARSP) vs Individual Lower Extremity Amputation Traditional Prosthesis (ILEATrad) vs. Control) will be considered as a between-subjects independent variable, Limb (prosthetic/intact or left/right) as a within-subjects independent variable and Speed (2.5, 3.0, 3.5, 4.0, 5.0, and 6.0 m/s) as a within-subjects independent variable. A two factor (Group x Speed) ANOVA will be used to identify statistical changes of dependent variables including ground reaction force symmetry, joint kinetic symmetry, and lumbopelvic kinematics. A one-way ANOVA will determine statistical differences between groups for the 50m dash time and speed. Significance for all statistical tests will be set at α=0.05. Sample Size Calculations This study will have appropriate statistical power to assess our hypotheses. Effect sizes for significant differences between RSP, intact, and able-bodied limbs in GRF loading rates, average GRF, and total joint work were estimated based on prior data collected by the PI from 8 ILEA and Control subjects running at 2.5, 3.0, and 3.5 m/s. These effect sizes ranged between 0.4 and greater than 0.9. An effect size of 0.7 was also estimated from the volume of oxygen (VO2) values of one study comparing subjects running with RSPs and traditional prostheses. To be conservative, the investigators utilized the lowest of these effect sizes (0.4) to perform a power analysis for ANOVA with repeated measures using G\*Power (v. 3.1.9.2, Dusseldorf, Germany). This yielded a total sample size of 24 with α = 0.05 resulting in 80% power to detect differences between limbs, prosthetic feet, or ILEA and Control subjects in kinetic values during running. The investigators anticipate similar effect sizes for changes in the remaining biomechanical variables the investigators will examine in the proposed study. Based on this analysis, 12 ILEA subjects and 12 control subjects will provide substantial statistical power to detect significant differences in biomechanical outcomes between groups and prosthetic feet.