Loading characteristics data applied on osseointegrated implant by transfemoral bone-anchored prostheses fitted with state-of-the-art components during daily activities

The data in this paper are related to the research article entitled “Load applied on osseointegrated implant by transfemoral bone-anchored prostheses fitted with state-of-the-art prosthetic components” (Frossard et al. Clinical Biomechanics, 89 (2021) 105457. DOI: 10.1016/j.clinbiomech.2021.105457). This article contains the overall and individual loading characteristics applied on transfemoral press-fit osseointegrated implant generated by bone-anchored prostheses fitted with state-of-the-art components during daily activities (i.e., microprocessor-controlled Rheo Knee XC knee, energy-storing-and-returning Pro-Flex XC or LP feet (ÖSSUR, Iceland)). Confounders of the loads are presented. The load profiles are characterized by the loading patterns, loading boundaries and loading local extrema of the forces and moments applied during straight-level walking, ascending and descending ramp and stairs at self-selected comfortable pace. The confounders of the loading information as well as new insights into inter-participants variability of loading patterns, loading boundaries and loading local extrema can inform the design of subsequent cross-sectional and longitudinal studies as well as literature reviews and meta-analyzes. The loading datasets are critical to clinicians and engineers designing finite element models of osseointegrated implants (e.g., medullar and percutaneous parts) and prosthetic components, algorithms capable to recognize the loading patterns applied on a residuum during daily activities, as well as clinical trials assessing the effects of particular prosthetic care interventions. Altogether, these datasets provide promoters of prosthetic care innovations with valuable insights informing the prescription of advanced prosthetic components to the growing population of individuals suffering from lower limb loss choosing bionics solutions. Online repository contains the files: https://data.mendeley.com/datasets/gmsyv97cpc/1

local extrema of the forces and moments applied during straight-level walking, ascending and descending ramp and stairs at self-selected comfortable pace. The confounders of the loading information as well as new insights into inter-participants variability of loading patterns, loading boundaries and loading local extrema can inform the design of subsequent cross-sectional and longitudinal studies as well as literature reviews and meta-analyzes. The loading datasets are critical to clinicians and engineers designing finite element models of osseointegrated implants (e.g., medullar and percutaneous parts) and prosthetic components, algorithms capable to recognize the loading patterns applied on a residuum during daily activities, as well as clinical trials assessing the effects of particular prosthetic care interventions. Altogether, these datasets provide promoters of prosthetic care innovations with valuable insights informing the prescription of advanced prosthetic components to the growing population of individuals suffering from lower limb loss choosing bionics solutions. Online repository contains the files: https://data.mendeley.com/datasets/gmsyv97cpc/1 © 2022 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ) Specifications Table   Subject Biomedical Engineering Specific subject area Design of prosthesis for individuals fitted with osseointegrated implant Type of data Table, graph How data were acquired Thirteen participants ambulated with an instrumented bone-anchored prosthesis made of tube and/or offset connector, transducer, Rheo Knee XC, Pro-Flex XC or LP feet (ÖSSUR, Iceland) and their own footwear. The tri-axial transducer measured directly and sent the loading data wirelessly to laptop nearby. Data format Raw, analyzed Parameters for data collection The forces and moments applied on and around the mediolateral, anteroposterior and long axes of transfemoral osseointegrated implant were recorded with sampling frequency set at 200 Hz and an accuracy better than 1 N and 1 Nm, respectively. Description of data collection Participants with transfemoral amputation conducted up to five trials of standardized daily activities (e.g., level walking, ascending and descending stairs and ramp) at self-selected speed using a instrumented bone-anchored prostheses.

Value of the Data
• The loading profile applied on transfemoral osseointegrated implants by bone-anchored prostheses fitted with state-of-the-art prosthetic components presented here were characterized by several datasets including the confounders as well as the loading patterns, loading boundaries and loading local extrema of the forces and moments applied during straight-level walking, ascending and descending ramp and stairs [2][3][4][5] . • These datasets are essential for promoters of prosthetic care innovations (e.g., users, clinicians, engineers, scientists, administrators) because they provide valuable insights supporting the prescription of advanced prosthetic components to the growing population of individuals suffering from lower limb loss choosing bionics solutions [6][7][8][9][10][11] . • The confounders of the loading information as well as the new evidence of inter-participants variability of loading patterns, loading boundaries and loading local extrema are required to inform providers of prosthetic care who will design subsequent cross-sectional and longitudinal studies (e.g., statistical planning, power calculation) as well as subsequent literature reviews and meta-analyzes [12][13][14] . • More precisely, the loading datasets are critical to clinicians (e.g., rehabilitation specialists) and engineers (e.g., manufacturers of components) designing finite element models of prosthetic components and osseointegrated implants parts (e.g., medullar and percutaneous parts), algorithms capable to recognize the loading patterns applied on a residual limb during daily activities, as well as clinical trials testing effects of particular interventions (e.g., design-based selection of components, alignment of prostheses) [15][16][17] .

Data Description
The confounders of the loading characteristics data including the selection criteria as well as the demographics, amputation, and residuum information as well as prosthesis and alignment of transducer are presented in Tables 1-7 and Fig. 1 , respectively. The confounders of the study design including non-experimental setup and number of gait cycles analyzed information are presented in Tables 8 and 9 , respectively.
The loading boundaries corresponding to the overall minimum and maximum of forces and moments applied on the implant expressed in units and percentage of the bodyweight were presented in Table 10 .
The mean and standard deviation of the pattern as well as the dispersion and mean for up to three local extrema of forces and moments during walking, ascending and descending ramp and stairs are presented in Figs. 2 , 4 , 6 , 8 , and 10 , respectively.
The box plots of magnitude of up to three local extrema of forces and moments during walking, ascending and descending ramp and stairs are presented in Figs. 3 , 5 , 7 , 9 and 11 , respectively.
The Mendeley Data include a spreadsheet and a report providing the confounders (e.g., selection criteria, demographics, individual amputation and residuum information, individual prosthesis and alignment of transducer data, description of non-experimental setup, number of gait cycles) and overall loading boundaries (e.g., minimum and maximum of forces and moments) of the loading data during level walking, ascending and descending ramp and stairs. Table 1 Selection criteria including inclusion and exclusion criteria applied for the recruitment and selection of participants using unilateral transfemoral bone-anchored prosthesis fitted with state-of-the-art components.

Inclusion criteria
1. To be fitted with osseointegrated fixation more than 6 months prior testing. 2. To be fully rehabilitated. 3. To have a clearance of at least 6 cm between percutaneous part of the fixation (e.g., abutment, dual cone) and prosthetic knee joint to fit the transducer. 4. To be able to be fitted with one of the nominated ÖSSUR components. 5. To be willing to participate to this project of research. 6. To be willing to comply with protocol. 7. To be able to walk 200 m independently with prosthesis. 8. To be between 18-80 years of age. 9. To be free of infection on the day of the recording session.    Table 4 Individual connection between the percutaneous part of the osseointegrated implant including or not a tube and/or an offset adapter (i.e., no tube and no adapter: 15%, a tube and an adapter: 8%, a tube and no adapter: 8%, no tube and an adapter: 69%) and the usual knee (i.e., N/A: 31%, Rheo Knee, OSSUR: 8%, Genium  ( continued on next page )   Table 7 Individual position of the distal end of the percutaneous part and the center of the Rheo Knee in relation to the origin of iPecsLab's transducer (RTC Electronics, USA) on the antero-posterior (AP), medio-lateral (ML) and vertical (VT) axes of the front and side views in the image (ICS) and transducer (TCS) coordinate systems.    Table 9 Breakdown of number of gait cycles analyzed for the cohorts of participants fitted advanced state-of-the-art components during five activities of daily living. Overall loading data     Fig. 11. Box plots showing low and high 95% confidence interval, mean and outliers of the magnitude of up to three local extrema (PT1, PT2, PT3) of forces and moments applied with state-of-the-art components during descending stairs.

Design
The study was designed as cross-sectional cohort study.

Participants
Thirteen participants with a single above-knee amputation fitted with press-fit osseointegrated implant participated in this study ( Tables 1 -3 ). They ambulated with a bone-anchored prosthesis. We estimated that this group corresponded to approximately 1.3% of the population of individuals with transfemoral amputation fitted with bone-anchored prostheses at the time of the recording, worldwide.
The Rheo Knee XC is a microprocessor-controlled knee. The Pro-Flex XC or LP feet are energystoring-and-returning feet. These components are referred to as "state-of-the-art". All Rheo Knee XC, Pro-Flex XC or LP feet are amongst the components frequently prescribed to individuals with osseointegrated implant worldwide, particularly in Australia [19] .
The tri-axial transducer of the iPecsLab was inserted between the participant's offset adapter and knee unit. It measured load data at sampling frequency set at 200 Hz and sent the data wirelessly to a laptop close by ( Tables 4 and 5 ). Forces and moments applied on mediolateral, anteroposterior and long axes of the implant were measured directly with an accuracy better than 1 N and 1 Nm, respectively [ 15 , 20 , 21 ].
All the percutaneous and medullar parts of the implant and the tube and/or connector were considered as a single rigid part. Nonetheless, the co-linearity of both long axes of the implant and the transducer varied according to the offset of the adapter ( Tables 6 and 7 , Fig. 1 ,).

Recording
Participants conducted a maximum of five trials of standardised daily activities, namely straight-line level walking, ascending and descending ramp and stairs ( Table 8 ) [ 5 , 22 ]. Participants were asked to complete each activity at a self-selected speed. They could use the handrails. Sufficient rest between trials was allowed to avoid fatigue when required.

Loading characteristics
The raw load data (e.g., forces and moments) recorded by the transducer were imported and processed into a specifically designed Matlab program (The MathWorks, Inc, USA) [ 4 , 16 ].
The load data was extracted through the following steps: 1. Calibration. The raw data were offset depending on the magnitude of the load recorded during calibration recording. 2. Detection of relevant segment. The first and the last strides recorded were eliminated to avoid the effects of gait initiation and termination so that the analysis included only steps taken at a steady pace.
3. Detection of gait events. Each heel contact and toe-off event was detected manually using loading profile applied on the long axis. 4. Time normalization. Loading data were time-normalization from 0 to 100 throughout the support phase. 5. Bodyweight normalization. Loading data were expressed as percentage of bodyweight [4] .
More advanced processing was required to characterize loading profile for each activity. This included extraction of loading patterns, loading boundaries (e.g., minimum and maximum of loading data across all gait cycles independently of the onset) and no more than three loading local extrema (e.g., onsets (%SUP) and magnitudes (%BW or %BWm) of points of inflection between loading slopes occurring consistently over successive gait cycles across all trials detected semi-automatically [ 1 , 4 ].
A loading pattern was described by its mean and one standard deviation. We reported confidence intervals calculated using the CONFIDENCE function in Microsoft Excel 2010 and the box plot showing low and high 95% confidence interval, mean and outliers created using SigmaPlot 11 (Systat Software, Inc, USA) for all discrete datasets (e.g., loading boundaries, local extrema) [14] .

Data availability
Loading data applied on osseointegrated implant by transfemoral bone-anchored prostheses fitted with state-of-the-art components: confounders and loading boundaries (Original data) (Mendeley Data).

Ethics Statement
Each participant signed a written ethical consent form approved by research organization's human ethics committee (Human Research Ethics Committee Certificate No 160 0 0 0 0332, Queensland University of Technology, Brisbane, Australia).

Transparency Document. Supporting Information
The data provided in Tables 1-4 , 6 , 8-10 in this article can be found in the online version at https://data.mendeley.com/datasets/gmsyv97cpc/1

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.