A comparison between three activity monitors in detecting commonly performed activities and postures of daily living

Abstracts / Osteoarthritis and Cartilage 24 (2016) S63eS534

osteoarthritis (knee OA), aim of this study is to analyze knee-joint motion and bone movements of femur and tibia relative to walking direction during stance phase of gait in the patients with knee OA of various severities. Methods: A total of 69 knees (62 adults) were divided into four medial knee OA groups using a Kellgren and Lowrence grading scale (grade-0: 12 knees, -I: 14 knees, -II: 16 knees, -III: 20 knees and -IV: 7 knees). Each subject’s gait was measured by a motion capture system (Sampling rate 120Hz), and bone positions of femur and tibia relative to ground were estimated by using a bi-planar X-ray and three-dimensional CT model of the bone. From continuous value of each bone position during stance phase of gait, knee kinematics was evaluated as tibial motion relative to femur. Additionally, the three-dimensional poses of femur and tibia relative to walking direction during stance phase were also calculated and these rotation angles to walking direction were evaluated. Results: Knee flexion range was significantly decreased with knee OA progression, whereas knee varus range tended to be increased. Femur and tibia at heel contact were positioned at external rotation to walking direction with knee OA progression. During loading response, the peak of femoral internal rotation relative to tibia was appeared in normal knee and early knee OA, whereas disappeared in severe knee OAs. The bone rotation ranges during stance phase were also decreased with knee OA progression. Conclusions: The increase of knee varus and the external rotations of femur and tibia at heel contact in severe knee OAs may be caused by the articular cartilage wear of knee joint and the external torsion of the long bone, and consequently it was considered that the bone rotation patterns during stance phase was change with knee OA progression. These findings suggest that knee kinematics during stance phase may affected by bone deformities of femur and tibia with knee OA progression, and thus it was important to clarify the effect between bone morphology and knee kinematics. 200 PATIENT-ORIENTED OUTCOME MEASURE FOR KNEE OSTEOARTHRITIS IS ASSOCIATED WITH GAIT ANALYSIS DATA OBTAINED FROM THE NOVEL DOWNSIZED MOTION CAPTURE TECHNOLOGY IN PATIENTS WITH THE END-STAGE KNEE OSTEOARTHRITIS H. Arita y, H. Kaneko y, M. Ishibashi z, R. Sadatsuki y, x, L. Lizu y, x, S. Hada y, M. Kinoshita y, J. Shiozawa y, Y. Takazawa y, H. Ikeda y, K. Kaneko y, x, M. Ishijima y, z. y Dept. of Orthopedics and Motor Organ, Juntendo Univ. Graduate Sch. of Med., Tokyo, Japan; z COI Program, Juntendo Univ. Graduate Sch. of Med., Tokyo, Japan; x Dept. of Sportology Ctr., Juntendo Univ. Graduate Sch. of Med., Tokyo, Japan Purpose: Osteoarthritis (OA) of the knee is an age-related progressive joint disease, which induces pain and gait disorder, resulting in a significant functional impairment, a loss of mobility and a diminished activity in daily life. While it has been revealed that the patient-oriented outcome measure in patients with knee OA reflects the clinical manifestation of the patients, such as pain severity, activity of daily living and social activities, it is still remained unclear whether the patient-oriented outcome measure is associated with data related to walking obtained from a gait analysis. In the present study we examined whether there was any association between the walking-related parameters using novel downsized motion capture technology and the patient-oriented outcome measure in patients with end-stage knee OA who decided to receive the total knee arthroplasty (TKA). Methods: This cross-sectional study was approved by the ethics committee of our university and was conducted between July 2014 and August 2015. Twenty-two patients with medial end-stage knee OA (Kellgren-Lawrence (K-L) grade 3 or 4) who were decided to receive TKA and were able to walk for thirty-meter or more were enrolled in the present study (male/female: 6/16, 73.8 y on average). The gait analysis was conducted using data obtained by the walking for thirty-meter, the one-leg standing with vision and the timed up and go test (TUG) using novel downsized motion capture technology consisting of linked stays (joint-angle measurement devices) and insole-type pressure sensor arrays (foot-pressure measurement devices). The patient-oriented outcome measure for knee OA used in this study was Japanese Knee Osteoarthritis Measure (JKOM), which is a patient-based, self-answering evaluation score that includes of four subcategories: pain and stiffness (JKOM II), activities of daily living (JKOM III), social activities (JKOM IV), and general health conditions (JKOM V) with 100 points as

S127

the maximum score. This measure has proven to have sufficient reliability and validity by means of statistical evaluation and comparison with the Western Ontario and McMaster Universities Arthritis Index (WOMAC) and the Medical Outcomes Study Short-Form 36-Item Health Survey (SF-36). Results: The gait speed of the patients were associated with JKOM total score (r ¼ 0.46, p ¼ 0.03), JKOM II (r ¼ 0.44, p ¼ 0.04), JKOM III (r ¼ 0.43, p ¼ 0.05) and JKOM IV (r ¼ 0.46, p ¼ 0.03), respectively. The TUG was associated with the JKOM total score (r ¼ 0.62, p & 0.01), JKOM II (r ¼ 0.63, p & 0.01) and JKOM III (r ¼ 0.50, p ¼ 0.02). The step length was inversely correlated with JKOM total score (r ¼ 0.61, p & 0.01), JKOM II (r ¼ 0.59, p & 0.01), JKOM III (r ¼ 0.55, p ¼ 0.01) and JKOM IV (r ¼ 0.58, p <¼ 0.01). The knee flexion angle while in walking was inversely associated with the JKOM total score (r ¼ 0.47, p ¼ 0.03), JKOM II (r ¼ 0.57, p ¼ 0.01) and JKOM III (r ¼ 0.43, p ¼ 0.05). The hip joint motion angle with sagittal plane while in walking was negatively correlated with JKOM total score (r ¼ 0.50, p ¼ 0.02), JKOM II (r ¼ 0.46, p ¼ 0.03), JKOM III (r ¼ 0.47, p ¼ 0.03) and JKOM IV (r ¼ 0.49, p ¼ 0.02). Pain VAS and JKOM V scores were not correlated with any gait parameters. Conclusions: Using a patient-oriented outcome measure in clinical medicine is important to put “evidence based medicine” into practice for the treatment of knee OA. On the other hand, it has been recently reported that self-report evaluation cannot evaluate correctively the improvement of walking disabilities and ADL limitations. The present study indicated that the most of the patient-oriented outcome measure for knee OA and its subcategories were associated with the walkingrelated parameters in knee OA patients, at least in patients with endstage knee OA. 201 A COMPARISON BETWEEN THREE ACTIVITY MONITORS IN DETECTING COMMONLY PERFORMED ACTIVITIES AND POSTURES OF DAILY LIVING M. Sliepen, D. Rosenbaum. Univ. sklinikum Münster, Münster, Germany Purpose: Knee osteoarthritis (KOA) is regarded as the foremost cause of functional limitations during daily life in elderly, such as difficulty in stair climbing. Being able to accurately monitor KOA patients’ activity during daily life could provide valuable information. Currently used activity monitors (AM) are only able to detect stepping and sedentary behaviour. Therefore, the aim of this study was to analyse whether a newly developed activity monitor can accurately classify additional activities and to compare its accuracy versus two established devices. Methods: Ten healthy subjects participated in this study (age 52.6 ± 19.8 yrs., BMI 27.6 ± 6.2 kg/m2). They were asked to wear three accelerometers simultaneously, while performing a specified protocol:  The Axivity3 (AX3, Axivity, UK) worn laterally on the thigh.  The ActivPal3m (AP3m, PalTechnologies Ltd., UK) fixed to the frontal aspect of the thigh.  The Move-II (KIT, Germany), worn on the pants/belt at the right anterior axillary line. The devices were worn according to the manufacturer’s specifications. The protocol contained the following activity classes, varying in distance or duration:  Walking (12.5e172.5 m)  Standing (10e30 s)  Sitting (30e60 s)  Cycling (345 m)  Stair ascending and descending (15-flight staircase) In order to unobtrusively film the activities, the participants wore a Hero 3 camera (GoPro Inc., USA) mounted on their chest. The protocol was designed to simulate a free-living environment. Therefore, none of the movements were performed in a laboratory setting. The mean percent error was calculated for the activity bout durations and the detected steps, compared to the video. Intra-class correlations (ICC) between direct observation and activity monitor output were determined per activity class. These analyses were performed for the AX3 and the AP3m. For the Move-II only the step detection accuracy was included for further analysis, due to poor classification results. P-values less than 0.05 were considered significant. Results: The AX3 accurately detected and correctly classified all activities (100%). The AP3m correctly classified sitting and standing behaviour (100%) but overestimated the amount of walking bouts with z170%. When examining the duration of each activity/posture class,

S128

Abstracts / Osteoarthritis and Cartilage 24 (2016) S63eS534

the AX3 and AP3m performed accurately, with minor errors (±5%). However, detecting the time spent descending stairs introduced higher errors (11.3% and 11.2% for the AX3 and AP3m respectively). The AP3m produced high, significant ICCs for every activity, whereas the AX3 produced moderate, yet significant ICCs for stair ascending and sitting (Table 1). All devices correctly estimated the amount of steps. Detecting the amount of cyclic rotations or stairs climbed produced larger errors of z10% and led to moderate ICC’s in several occasions (Table 2). Conclusions: The AX3 showed promising results in detecting activity classes. Furthermore, the AX3 detected the time spent within these classes and the amount of steps with reasonable accuracy (at least 90%). The AP3m achieved marginally more accurate results and attained higher ICCs in general. The Move-II detected steps slightly more precisely, but failed to correctly classify activities. It should be noted that the AP3m and the Move-II do not (accurately) detect cycling and stair climbing. In conclusion, since the AX3 is able to detect a larger variety of activities with reasonable accuracies, it provides more detailed information on a patient’s physical activity during daily life.

Table 1 Mean percent error (±SD) and ICC of the time spent within a specific activity class AP3m

AX3

Walking Standing Sitting Cycling Stair ascending Stair descending

Mean ± SD

ICC

Mean ± SD

ICC

3.3 ± 2.7 0.1 ± 6.5 4.2 ± 3.1 4.1 ± 6.4 4.1 ± 18.1 11.3 ± 15.4

0.99* 0.87* 0.56* 0.94* 0.54* 0.90*

3.8 ± 2.5 3.2 ± 6.8 0.2 ± 1.0 4.2 ± 6.1z 7.2 ± 12.4z 11.2 ± 9.0z

0.97* 0.81* 0.98* 0.94* 0.85* 0.93*

*p < 0.05, zClassified as ‘Stepping behaviour’ by the AP3m

stair ascent and descent although no significant differences were found for KAM. This study presents case studies of subjects with lateral knee OA in which medial and lateral knee contact forces (KCF), accounting for the combined effect of intersegmental, musculotendon and ligament forces, are calculated while stair ascending and descending. We hypothesize that subjects with lateral knee OA will present reduced loading on the lateral compartment trying to decrease the pain, particularly during descending stairs. Methods: Four patients with knee OA at lateral compartment (mean age of 66.5 ± 5.4 and BMI of 28.0 ± 4.3 kg/m2) were recruited. Patient classification was based on Magnetic Resonance Imaging (MRI) demonstrating articular cartilage degeneration, subchondral BMLs and presence of osteophytes (Table 1). Hip and Knee disability and Osteoarthritis Outcome Score questionnaires assessed knee and hip pain. Nine healthy subjects, with no signs of knee OA as determined using MRI or complaint of pain, have also participated (mean age of 49.7 ± 13.2 and BMI of 27.4 ± 3.7 kg/m2). Motion analysis was performed while ascending and descending a staircase consisting of seven steps at self selected speed. A 10-camera 3D motion capture system (Vicon) synchronized with four force platforms (embedded in the middle of the staircase) recorded the 3D position of 34 reflective markers according to an extended Helen Hayes protocol, at 100 Hz and measured ground reaction forces at 1,000 Hz (Kistler). A multi-body right knee model with 6 DoF for each tibiofemoral and patellofemoral joints was used. The 3D model consisted of 14 ligament bundles as nonlinear elastic springs. Cartilage contact pressures and forces were computed using an elastic foundation model. The knee model was incorporated into a generic multibody lower extremity musculoskeletal model that includes 44 muscle-tendon units. An enhanced static optimization routine was used to calculate the muscle forces. KCF were normalized to body weight (BW). Results: Patients with lateral knee OA reduced first peaks KFM and KAM while ascending stairs (Figure 1). Decreased first peak KCF were found at the lateral (0.807 ± 0.127 xBW) and medial (1.712 ± 0.499 xBW) compartments (Table 2). However, during the second half of the stance phase, these patients showed higher KCF on both compartments. For descending stairs (Figure 2), reduced peak KFM but increased KAM were observed during the first peak. A clear reduction of lateral compartment KCF compared to controls was found, except at the first peak.

Table 2 Mean percent error (±SD) and ICC of the step detection within a specific activity class AP3m

AX3

Walking Cycling Stair ascending Stair descending

Move-II

Mean ± SD

ICC

Mean ± SD

ICC

Mean ± SD

ICC

1.4 ± 1.9 11.2 ± 26.1 7.4 ± 10.5 12.6 ± 14.5

0.99* 0.47 0.67* 0.41

1.6 ± 1.4 6.4 ± 4.1z 4.0 ± 7.3z 14.1 ± 19.7z

1.00* 0.95* 0.78* 0.21

0.0 ± 4.0 83.3 ± 37.3z 3.7 ± 12.7z 5.6 ± 7.5z

0.97* 0.249 0.57* 0.83*

*p < 0.05, zClassified as ‘Stepping behaviour’ or walking by the AP3m and Move-II, resp.

202 ALTERED KNEE LOADING IN PATIENTS WITH LATERAL KNEE OA WHILE ASCENDING AND DESCENDINS STAIRS e CASE STUDIES S.P. Ferreira Meireles y, N.D. Reeves z, C.R. Smith x, D.G. Thelen x, R. Hodgson k, I. Jonkers y. y KU Leuven, Leuven, Belgium; z Manchester Metropolitan Univ., Manchester, United Kingdom; x Univ. of WisconsinMadison, Madison, WI, USA; k The Univ. of Manchester, Manchester, United Kingdom Purpose: Subjects with knee osteoarthritis (OA) often complain first of pain during weight-bearing activities involving bending of the knee, as climbing or descending stairs. Indications of lower external flexion moments (KFM) and lower external adduction moments (KAM) during

Figure 1. Average KAM and KFM, and KCF during stance phase while ascending stairs.