Seventeen female college badminton players with no history of knee injury participated in this study. The average age was 20.3 ± 1.8 years, height was 158.4 ± 5.4 cm, body weight was 51.7 ± 5.8 kg and athletic experience of badminton was 5.3 ± 2.1 years. All subjects were right hand dominant, based on racket holding preference. No subject had pathological complaints. Administered orthopedic physical examinations revealed no loss of range of motion or muscular weakness in trunk, lumbar spine, hip, knee or ankle joint.

Trial data were collected with a 3-dimensional motion analysis system (Vicon; Oxford Metrics, Oxford, UK) utilizing seven infrared cameras. Kinematics data were sampled at 120 Hz and recorded digitally on a Dual Pentium III 1 GHz personal computer. Cameras were positioned so that each retroreflective marker was detected by at least two cameras throughout the task. Ground reaction force was collected at 120 Hz using a calibrated and leveled force plate (model OR6-6-1; AMTI, Watertown, Massachusetts, USA) embedded in the floor and synchronized with a Vicon system for simultaneous collection. Each subject was barefooted and wore black shorts during the testing. Sixteen retroreflective markers (25 mm) were placed on specific anatomical landmarks (anterior superior iliac spine, posterior superior iliac spine, mid thigh, lateral condyle of the femur, mid lower leg, lateral malleolus of ankle, heel and second metatarsalis) to calculate motion of the hips, knees and ankles in the sagittal, frontal and transverse planes (Kadaba et al. 1990). Three dimensional marker trajectories were recorded and kinematic and kinetic variables were calculated using Plug-in Gait in Vicon Workstation (version 4.6; Oxford Metrics, Oxford, UK). Inverse dynamics analyses were used to calculate external joint moments from kinematic data and force plate data with Plug-in Gait. Marker trajectory data were filtered using a Woltering quintic spline filter with a predicted mean square error of 20 mm (Woltering 1986). The force plate data were filtered through a low-pass Butterworth digital filter at a cut-off frequency of 12 Hz. All kinetic data were normalized to body weight (kg) and height (m).

Before participation, the testing procedures were explained to each subject. Anthropometric measurements including height, weight, segmental lengths and diameters of the ankles and knees, length of the feet, and pelvic width were recorded for each subject. The footwork and overhead stroke task were then demonstrated to each of the subjects by a badminton coach who had approximately 15 years of experience as a competitive player. Subjects were asked to perform two different testing tasks: a right back-step task and a left back-step task. In the right back-step task, the subjects stepped back three steps 45° diagonally to the right from the starting position. This simulated the back-step towards the forehand-side rear of the badminton court. They then made an overhead stroke with a racket in their right hand, immediately landed with their left leg on the force plate and then returned to the starting position (Video 19.3). In the left back-step task, the subjects stepped back three steps 45° diagonally to the left from the starting position. This simulated the back-step towards the backhand-side rear of the badminton court. This was followed by an overhead stroke. They then landed on the force plate and returned to the starting position in the same way as in the right back-step task (Video 19.3). Subjects were allowed to practice the tasks several times, and then performed three to five consecutive trials. To avoid any coaching effect on the subjects’ natural performance of the tasks, no further instructions relating to stepping, landing or stroking techniques were provided.

The data analysis was performed using a trial in which the subject successfully landed within the force plate without overrunning it. In general, right-handed badminton players land on the left leg after an overhead stroke (Video 19.1). Therefore all data were analyzed for the left leg. Because non-contact ACL injuries generally occur in the initial phase of landing (Koga et al. 2010), the data analysis was only focused on the impact phase after an overhead stroke. The impact phase was defined as between the time of initial contact (IC) of the left toe and the floor after an overhead stroke and maximum knee flexion (MKF). All kinematic data were time normalized to the impact phase. The values thus ranging from 0 (equal to IC) to 100 % (equal to MKF). From each trial, the flexion/extension, adduction/abduction and internal/external rotation angles of the left hip, the flexion/extension and varus/valgus of the left knee, and the dorsi/plantar flexion, inversion/eversion and adduction/abduction angles of the left ankle at IC and MKF were determined. The peak values of the external knee valgus moment were extracted from IC to MKF, and normalized to height and body weight. The mean and SD of the relevant measures were calculated for all trials. To evaluate the significance of the comparison between the right and left back-step tasks for kinematic and kinetic data during single-leg-landing, paired t tests were utilized. A significance level of 0.05 was adopted. Statistical analyses were conducted using SPSS Version 16.0 software (SPSS Inc, Chicago, Illinois, USA).