The reliability of lower limb 3D gait analysis variables during a change of direction to 90- and 135-degree manoeuvres in recreational soccer players

To be enrolled in the study, a participant must be of the general population, healthy, and physically active. The participant must be a recreational, non-elite soccer player who practices soccer for at least 30 minutes three times a week over the last six months in regular basis. Moreover, each participant had to practice 90 ∘ and 135 ∘ COD manoeuvres in their routine sports activity. Participantion was limited to those between 18 and 35 years old, since they practice soccer the most and are the most prone to injury [29]. Those with previous injuries in the last six months were excluded. An injury was defined as any musculoskeletal complaint that led them to stop their regular exercise activity. Participants were excluded if they were overweight (above 24.9 BMI), had any deformities or disease known to affect their ability to walk and run, or were not able to give informed consent. Any individual who was not able to follow the procedure was not allowed to participate in the study.

2.2 Procedure

A Qualisys motion analysis system (Gothenburg, Sweden) with ten cameras (Qualisys Oqus 700+) synchronized with three force platforms (AMTI BP400600, USA) operating via Qualisys Track Manager software (version 2.16) was used. The sampling rate was 250 Hz for kinematics and 1000 Hz for kinetics. Participants were tested twice at the same time of the day, one week apart. The selection of a time interval in reliability depends on whether the time should be enough to reduce recall bias and not too long to cause real change. Based on previous similar studies, seven-day intervals were selected [20, 30].

Before the participants arrived in the lab, the lab was calibrated. A supervisor familiar with specialised techniques manages the lab and calibrates the force platform to ensure that it runs perfectly regularly. The calibration process starts by placing the L-shaped metal frame in a previously specified place along the corner of one of the force platforms. Then, the wand is waved randomly in the required volume. For the calibration to be accepted, it must get residual volume below 1 mm for each camera based on the manufacturer’s recommendation.

Upon the participants’ arrival, the experiment was described, and consent forms were distributed and obtained after providing enough time for participants to think, ask questions, and decide. Demographic characteristics (age, mass, height) and previous medical histories were taken. Then, each participant was asked to change into shorts and a T-shirt. Standardized shoes (New Balance, UK) were used to reduce the possibility of interaction between shoes and the surface. The Calibrated Anatomical System Technique (CAST) method was used to place the markers [31], as shown to reduce error compared to an earlier model [32]. The CAST model allows for two sets of 14.4 mm markers (technical marker, anatomical marker). The anatomical marker is used to define the local coordinate system in relation to the anatomical frame, while the technical marker is used to track movement. The segment was defined by the proximal and distal endpoints (Table 1) [33]. Four clusters were attached securely with a Velcro strap in each participant’s shank and thigh in an anterior lateral direction (Fig. 1). Each participant was given enough time to practice until they felt comfortable and natural. The static trial was captured, and after that, static markers were taken off [34].

OPEN IN VIEWER

To perform the required task, participants were asked to run in a straight direction for five meters, and when they hit the force platform by the required limb, they changed the direction (90 ∘ or 130 ∘ ) toward the opposite limb and ran for three meters. To guide the participants and ensure that all of them performed the same required angle, cones were placed along the track (Fig. 2). Contacting the force platform with the selected leg was achieved by monitoring the participants’ starting point in relation to different coulure taps on the floor. Five successful trials were conducted for each task and limb after performing five minutes of low-intensity warm-up (cycle ergometer) to help avoid any injury and reduce the risk of discomfort [35, 36]. A successful trial was defined as one in which the foot fully contacts the force platform with the required speed and good marker view. Rest times of five minutes between tasks and a half minute between trials were provided to reduce the effect of fatigue. A Brower Timing Gate system (TC-Timing System, USA) was used to control the speed and was placed along the eight-meter path at the hip level. The speed was controlled for 4.2 m/s ± 0.5 to allow comparison between limbs, tasks and previous literature [26]. For the current study, the selected limb in the analysis was assigned via Randomization.com. The same procedure was applied in the second test after seven days by the same examiner. To improve the reliability of the data, the distance between the markers and the floor and the distance between the markers were measured. These were used in the second session to improve accuracy.

OPEN IN VIEWER

Peak vertical ground reaction force (VGRF), peak knee flexion angle and peak abduction angle, peak external knee abduction moment (KAM), peak hip joint flexion, adduction, and internal rotation angle outcomes were selected. The rationale for choosing these outcomes is that higher VGRF will lead to higher KAM, and a previous study showed an increased risk in ACL with higher VGRF [13]. An increase in the knee abduction angle has been linked to an increased risk of ACL injury and KAM increase [16, 37]. In the sagittal plane, reducing the knee flexion angle has been linked to increasing ACL loads [13]. A higher KAM leads to an increase in the tension on the ACL and an increase in the risk of injury [16]. A previous review highlighted that the sagittal plane hip had been linked to the occurrence of ACL injury [16]. In contrast, transverse plan hip motion has been linked to increased abduction through dynamic valgus [16]. Moreover, the increase in the hip adduction angle was found to be a significant predictor for the knee abduction angle [11].

The raw data were captured and labelled through Qualisys Track Manager Software (version 2.16). After labelling, each trial was exported as a visual 3D file to be processed in Visual 3D (Version 6.00.16, C-Motion Inc., Rockville, MD, USA). In Visual 3D, the kinematics and kinetics data were filtered by a 25 Hz and 12 Hz Butterworth fourth-order bi-directional low pass filter, respectively, and interpolated for ten frames. This filter cut-off was selected based on previous studies [38, 39].

The lower extremity model was then created and modelled as a conical frustra using the inertial parameters estimated via the anthropometrics data. X-Y-Z Euler rotation sequences were used to process the joint kinematic angles, where X represents flexion-extension, Y means abduction-adduction, and Z represents internal-external rotation [40]. The joint kinematic data were calculated based on inverse dynamics theory. Joint moments were normalised on body mass and presented as an external moment, while kinetics and kinematics data were normalised on 100% of the stance phase. Initial contact was defined as the point when VGRF exceeds 20 newtons, while toes off when VGRF falls below 20 newtons [20].

2.5 Statistical analysis

The required sample was calculated based on a previous method published in 2018 [41]. The minimum accepted reliability value for ICC was 0.40 in the equation. The expected reliability value for the ICC was between 85 to 95 with 90% power, which shows that the required sample size ranged from 7 to 17 participants.

The statistical analysis was conducted using Statistical Package for Social Sciences (SPSS) software version 21. The mean of the five trials from both visits was used to calculate reliability. An ICC two-way mixed model with absolute agreement was used since only one investigator conducted all the measurements [42]. The ICC model was interpreted according to the following criteria: 0.40 to 0.70 fair, 0.70 to 0.90 good and 0.90 and above excellent [43]. The confidence interval (CI) and standard deviation (SD) were also calculated and presented. Moreover, the standard error of measurement (SEM) was calculated since the ICC cannot alone provide any indication of the level of disagreement [23]. SEM was calculated based on the following formula: SD* SQR(1-ICC) [44]. SQR can be defined as a square root. SEM provides a number with the same unit for the outcome measure, with a lower value indicating low measurement error. The mean of both visits and absolute difference between visits were calculated.

3.1 The reliability of 90 ∘ COD manoeuvres

Table 2 represents the between-day reliability for 90 ∘ COD manoeuvres. In general, the ICC values ranged from 0.98 to 0.88, which was interpreted as good to excellent. The absolute difference between day one and day two for hip flexion, hip adduction, hip internal rotation, knee flexion, and knee abduction angles ranged from 1.6 ∘ to 4.2 ∘ . For the KAM, the difference between the first and second visits was low (0.10 Nm/Kg). The VGRF value changed slightly between visits with 0.09 body weight (BW). The SEM for all variables in 90 ∘ COD manoeuvres was low, with only up to 1.2 ∘ for angle, 0.03 Nm/Kg for KAM, and 0.01 BW for VGRF. In the 90 ∘ COD manoeuvres, hip flexion angle, hip adduction angle, hip internal rotation angle, knee flexion angle, knee abduction angle, KAM, and VGRF average between-day values were 48.4 ∘ , - 12.1 ∘ , 7.7 ∘ , 63.0 ∘ , - 6.4 ∘ , 0.825 Nm/Kg, and 2.285 *BW, respectively.

3.2 135 ∘ COD manoeuvres reliability

Table 3 represents the between-day reliability for 135 ∘ COD manoeuvres. In general, the ICC values ranged from 0.95 to 0.85, which was interpreted as good to excellent. The SEM of the joint angles was below 1.69 ∘ , representing a low measurement error, while the maximum absolute difference was up to 4.3 ∘ (for hip joint internal rotation angle). For KAM, the absolute difference between days was 0.14 Nm/kg with SEM of 0.10 Nm/kg. The VGRF showed low SEM with 0.05 and 0.19 BW absolute difference between days. In the 135 ∘ COD manoeuvres, hip flexion angle, hip adduction angle, hip internal rotation angle, knee flexion angle, knee abduction angle, KAM, and VGRF average between-day values were 51.4 ∘ , - 15.2 ∘ , 8.5 ∘ , 67.5 ∘ , - 7.0 ∘ , 0.83 Nm/Kg, and 2.16 *BW, respectively.