Data were collected during two sessions (baseline and six-week follow-up) in a university research laboratory and via REDCap (Research Electronic Data Capture, hosted by the University of Otago, Dunedin, New Zealand). Ethical approval was granted by the Health and Disability Ethics Committee (of New Zealand). We followed CONSORT reporting guidelines [15]. The sample of this study was the same sample as in the previous report [5].
Trial design and binding
Part 1 of the study consisted of a cross-over laboratory-based study, exploring immediate effects of wearing the knee sleeve. It was impossible to blind participants and assessors to the sleeve condition in Part 1. Part 2 was a parallel two-armed, assessor-blinded randomised clinical trial (RCT), with the same participants as in Part 1 [5]. The biostatistician and the research assistant were blinded to group allocation for the RCT.
Participants
Recruitment
We recruited participants via community advertising and using TrialFacts (https://trialfacts.com/), a research participant recruitment agency. Recruitment started in September 2018 and closed in September 2020 due to the end of the funding period. Volunteers completed a questionnaire (also serving as screening for eligibility) via REDCap prior to attending the first laboratory session [5]. The questionnaire included demographics, injury and surgery history, the International Knee Documentation Committee Subjective Knee Form (IKDC-SKF) [16] and the Tegner activity scale [17].
Inclusion criteria
We recruited men and women, aged 18–40 years, who underwent ACL reconstruction within 6 months to 5 years previously [5]. We targeted individuals who had not yet reached full functional level, for the purpose of this study defined by a score between 40 to 80/100 on the IKDC-SKF [16, 18, 19]. The initial protocol defined a time period for an ACL reconstruction as 6 months to 3 years previously. Due to slow recruitment rate, and amendment was approved by the ethics committee to expand the period until 5 years post-ACL reconstruction.
Exclusion criteria
We excluded participants who had undergone a revision ACL reconstruction of the same knee (due to re-injury), or a previous ACL reconstruction of the opposite knee; self-reported any other lower limb, pelvic or low back musculoskeletal injuries or disorders that required medical care over the past 6 months; had known systemic, neurological or cardiovascular disorders; or had a body mass index (BMI) above 30 kg/m2 [5]. We excluded participants with an IKDC-SKF score less than 40 (due to potential safety risk during the laboratory-based tasks) or greater than 80/100 (as use of a sleeve would clinically be less likely to add benefit) [5].
Procedures
Randomisation
We randomised each participant twice (once for the cross-over trial, and once for the RCT) with equal numbers in each group for both allocations. The research officer sequentially block randomised groups of 8 participants with an electronic random number generator before participants entered the study. Each group was stratified by sex. The research officer informed the assessor for the laboratory data collection of the order for the conditions for the cross-over trial, and the group allocation (for the RCT) via email prior to the participant’s first laboratory session [5]. Participants provided written informed consent at the start of the first session. Participants were dressed in a singlet, a pair of shorts and their own sport shoes. Body mass and height were measured during the baseline session [5].
Part 1: Laboratory cross-over trial
Participants undertook two hopping tasks: a maximum horizontal single leg hop and a sub-maximum step-down hop. A sub-maximal level was chosen for safety as we were seeking participants with remaining residual functional limitations (as defined by the IKDC-SKF). To allow comparison of knee mechanics across the six-week period, the distance of the required hop was standardised and individualised based on the participants single-leg hop distance of the uninjured side.
Participants practised the hopping tasks at sub-maximal distance with the uninjured and injured sides until they were confident with performing them as part of familiarisation and warm-up. They performed the maximum horizontal hop prior to undertaking the step-down-hop.
Part 2 Randomised clinical trial
On completion of the first laboratory session, the assessor informed participants of their group allocation for the RCT. All participants were asked to return to the laboratory following the 6-week period to repeat the above assessments, repeating the hopping tasks without wearing the knee sleeve [5].
Intervention
We used a commercially available knee sleeve, GenuTrain (Bauerfeind® AG, Zeulenroda-Triebes, Germany), a CE-certified medical device, as the intervention. The sleeve consisted of flexible elastic/knitted materials, providing knee support without limiting the range of motion [5]. All participants performed the step-down hop with and without the sleeve for the cross-over trial (Part 1). For Part 2 (RCT), only participants of the ‘Sleeve Group’ (intervention) were provided with the knee sleeve for the 6-week period. They were instructed to wear the sleeve for a minimum of 1 hour per day during their rehabilitative exercises, physical activity and sports. The control group were not provided with a sleeve during this period. As reported previously [5], participants of the ‘Sleeve Group’ were asked to document the use of the knee sleeve in a daily diary (Microsoft® Excel spreadsheet) and all participants were asked to document their daily physical activity and exercise.
The researcher explained the use of the knee sleeve to the ‘Sleeve Group’ and provided them with an instructional leaflet. They were informed to discontinue use of the knee sleeve and contact the researcher should any side-effects evolve, such as discomfort during use, pain, burning sensations of the knee, leg or foot, are swelling of the knee or calf [5].
Outcomes
1) Single-leg horizontal jump: After familiarisation with the tasks, they performed 3 trials of single-leg maximum horizontal jump, as described in the first report [5]. The maximum jump distance of the uninjured side during Session 1 was used to calculate the individualised target for the hop distance during the step-down task. The target step-down hop distance was 60–70% of the maximum single leg hop distance of the uninjured side, and was kept constant for the participant across the two sessions.
2) Step-down hop: Three-dimensional motion analysis was performed for the step-down hop with 11 infra-red cameras (Motion Analysis Corporation, Santa Rosa, CA, USA), sampling at 120 Hz, and Cortex 2.0 software, synchronized with a floor-mounted force plates (BP2436 AMTI Inc., Newton, MA, USA), sampling at 2,400 Hz. A set of 42 reflective markers (diameter 12.5 mm) were applied to the trunk, pelvis and lower extremities (Fig. 1). The marker positions were tracked by the camera system, reconstructed in 3D space and used to define segment coordinate systems.
Following the marker placement, a static anatomical calibration trial was performed with the participants standing still, on both feet. A functional movement trial was then performed by moving the hip in the three planes for calculation of functional hip joint centres [20]. One trial was undertaken for each side respectively.
The participants were then asked to stand on a 30-cm box, placed 15 cm from the force plate, and performed a step-down hop (adapted from Kristianslund and Krosshaug [21]) onto the force plate: the participant was asked to step off the box with either the injured or the uninjured leg onto the force plate (Fig. 2A), then hop forward off the plate as fast as possible. The distance of that hop was 60–70% of the maximum horizontal jump distance of the uninjured side (Fig. 2B). They performed the step-down-hop with the uninjured side first, then the injured side under the (1) ‘control’ condition (no sleeve) and (2) the ‘sleeve’ condition (experimental, wearing the sleeve), ordered by randomisation. A 5-minute walk between the conditions provided a standardised run-in to the second condition to minimise carryover effects.
Data processing and analysis
Cortex 5.5 (Motion Analysis Corporation, Santa Rosa, CA, USA) was used to track and label the markers, and the biomechanical model, kinematic (joint angles) and kinetic (moments) data were calculated using Visual3D Professional v6 (C-Motion, Inc., Germantown, MD, USA). Functional hip joint centres were estimated based on a movement trial involving all degrees of freedom of the hip [22], while anatomical joint centres at the knee and ankle joints were based on medial and lateral markers [20]. Marker clusters attached to a rigid base were fixed to the thighs and shanks to minimise marker coordinate estimation error [23]. Kinematic and kinetic data were filtered using a low-pass, double, second order Butterworth filter with a cut-off of 10 Hz.
Data from the stance phase of the hop were of interest; the start and end of the stance phase were defined by the vertical component of the ground reaction force exceeding and returning below 20 N, respectively. Peak knee flexion angles and flexion angles at initial contact were extracted for the stance phase. Based on the biomechanical model’s segment inertial properties and joint properties, inverse dynamics analyses were applied to the model kinematics and ground reaction forces to calculate the net joint moments. Net joint moments were resolved into the proximal segment’s coordinate system and represented as vectors. The moments were normalised to body size (Nm/BW*HT). The temporal variable was the duration of the stance phase of the step-down hop. For all variables, the averages of five trials for each limb (injured versus uninjured) and condition (sleeved and unsleeved) for each participant were calculated.
Sample size
We report a secondary analysis of a larger study investigating effects of wearing a knee sleeve for individuals following ACL reconstruction. The sample size was based on the primary outcome measure, the horizontal hop distance, as reported in [5]. We did not undertake a sample size calculation specifically for the outcomes reported in the current paper.
Data analysis
Demographic data were presented descriptively (means and standard deviations for approximately normally distributed continuous variables; medians and ranges for other continuous variables; and counts and percentages for categorical variables).
Statistical parametric mapping
We analysed kinetic and kinematic trajectories using Statistical Parametric Mapping (SPM, http://spm1d.org/; Pataky, 2012) [24]. SPM allows comprehensive statistical analyses of the trajectory of a given biomechanical variable (such as knee angles) during a specific task, using the entire n-dimensional biomechanical sampling data. That contrasts with discrete point analyses, such as peak knee flexion, that assess only the one point during the entire movement. SPM uses Random Field Theory to make probabilistic conclusions based on the random behaviour of that 1D observational unit. Scalar observations of discrete variables are based on traditional 0D Gaussian randomness [25]. Thus, the whole trajectory can be assessed with SPM, and hypotheses relating to differences at specific time points during the task (such as a hop) do not need to be defined a priori.
Stance phase kinematic and kinetic trajectories in the sagittal plane were temporally normalised to stance duration using linear interpolation across 100 equally spaced time points. Trajectories of five trials for each limb and each session were computed using MATLAB R2018b (The MathWorks Inc., Natick, MA, USA). The mean trajectory for the uninjured and injured sides (sleeved and unsleeved conditions, Session 1) were then computed.
For Part I, one-way repeated measures ANOVA was used to compare three conditions at the baseline: (a) uninjured side, unsleeved to (b) injured side, unsleeved, and (c) injured side, sleeved. ANOVA was conducted separately for each of two dependent variables: (i) knee flexion, and (ii) external flexion moment, and a conservative Bonferroni threshold of 0.025 was adopted to correct for multiple comparisons across these two tests. Post-hoc analyses using paired t tests and a secondary Bonferroni correction across the three pairwise tests (a + b, a + c and b + c) were used to assess between-condition effects.
For Part 2, within-participant changes were depicted between baseline and follow-up sessions (unsleeved only). An SPM independent t test was used to compare within-participant changes in the Sleeve group versus the Control Group. This t test was repeated for each of the two aforementioned dependent variables, again with a conservative Bonferroni correction of 0.025.
Discrete variable analysis
Post hoc analyses were also performed for the discrete variables. Analyses for Part 1 (cross-over trial) were adjusted for participant sex, surgery type, and time since surgery (as a continuous measure), and sequence effect. Analyses for the knee flexion angles (angle at initial contact, peak angle, and excursion), knee peak external flexion moment and stance duration are from linear mixed models using Restricted Maximum Likelihood (REML) to estimate random effects. A random participant effect and, where both limbs are measured at multiple time points, a random measurement occasion effect nested within participant. These analyses were performed with Stata (16.1, StataCorp LLC, College Station, TX, USA).
For Part 2 (RCT), individual change scores from baseline to follow-up were calculated for knee flexion angles, moments and stance duration. The change scores were compared between the Sleeve Group and the Control Group using Mann-Whitney U tests for each outcome. The alpha level was set at p ≤ 0.05. These analyses and those of demographic data were performed with SPSS Version 24.0 (IBM Corp, Armonk, NY). Test-retest reliability from our laboratory for discrete variables of peak knee flexion, external peak flexion moment and stance duration are shown in the supplementary file.