The Knee 22 (2015) 618–623
Contents lists available at ScienceDirect
The Knee
Patient-reported outcome measures versus inertial performance-based outcome measures: A prospective study in patients undergoing primary total knee arthroplasty S.A.A.N. Bolink ⁎, B. Grimm, I.C. Heyligers AHORSE Foundation, Dept. of Orthopaedics, Atrium Medical Center Heerlen, The Netherlands
a r t i c l e
i n f o
Article history: Received 10 December 2014 Received in revised form 11 March 2015 Accepted 15 April 2015 Keywords: Total knee arthroplasty Outcome assessment Performance-based test PROMs Inertial measurement unit
a b s t r a c t Background: Outcome assessment of total knee arthroplasty (TKA) by subjective patient reported outcome measures (PROMs) may not fully capture the functional (dis-)abilities of relevance. Objective performance-based outcome measures could provide distinct information. An ambulant inertial measurement unit (IMU) allows kinematic assessment of physical performance and could potentially be used for routine follow-up. Aim: To investigate the responsiveness of IMU measures in patients following TKA and compare outcomes with conventional PROMs. Methods: Patients with end stage knee OA (n = 20, m/f = 7/13; age = 67.4 standard deviation 7.7 years) were measured preoperatively and one year postoperatively. IMU measures were derived during gait, sit–stand transfers and block step-up transfers. PROMs were assessed by using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and Knee Society Score (KSS). Responsiveness was calculated by the effect size, correlations were calculated with Spearman's rho correlation coefficient. Results: One year after TKA, patients performed significantly better at gait, sit-to-stand transfers and block stepup transfers. Measures of time and kinematic IMU measures demonstrated significant improvements postoperatively for each performance-based test. The largest improvement was found in block step-up transfers (effect size = 0.56–1.20). WOMAC function score and KSS function score demonstrated moderate correlations (Spearman's rho = 0.45–0.74) with some of the physical performance-based measures pre- and postoperatively. Conclusion: To characterize the changes in physical function after TKA, PROMs could be supplemented by performance-based measures, assessing function during different activities and allowing kinematic characterization with an ambulant IMU. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Total knee arthroplasty (TKA) has developed into a successful intervention for patients with advanced knee osteoarthritis (OA). The incidence of knee OA and the number of TKAs are expected to increase substantially in the future [1] with a bigger proportion of relatively young patients who are more active and more demanding [2]. This challenges the outcome assessment methods we use, which equally need to adapt and evolve [3]. Traditionally, outcome assessment after TKA includes pain and function and it is important to have separate measures of both domains. Pain is often considered the most important aspect of outcome however it is a subjective measure and influenced by socioeconomic and psychosocial factors [4]. Function is becoming more important, especially for the new generation of physically more active patients, and can be assessed by either patient-reported measures ⁎ Corresponding author at: Dept. of Orthopaedics, Atrium Medical Center Heerlen, Henri Dunantstraat 5, 6419 PC Heerlen, The Netherlands. Tel.: +31 655341492. E-mail address: [email protected] (S.A.A.N. Bolink).
http://dx.doi.org/10.1016/j.knee.2015.04.002 0968-0160/© 2015 Elsevier B.V. All rights reserved.
(e.g. questionnaires) and performance-based measures (e.g. gait analysis) [5,6]. Popular patient-reported outcome measures (PROMs), such as the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and Knee Society Score (KSS), combine measures of pain and function and are convenient to use in longitudinal follow-up. However, PROMs are subjective measures, suffer from a ceiling effect and their pain dominance masks the functional changes [7]. Therefore, outcome assessment by PROMs only may not fully capture the functional (dis-)abilities of relevance. Performance-based outcome measures on the other hand tend to be more objective and measure what a patient actually can do [8]. As performance-based measures have a low correlation with PROMs, they may capture other dimensions of physical function and could provide distinct and supplementary information [5]. Thus, PROMs and performance-based tests are both essential components of comprehensive functional outcome assessment for longitudinal follow-up. Unfortunately, performance-based measures struggle with feasibility issues as they are mostly time-consuming and require dedicated laboratories. This justifies the frequently used, and easy to measure, timed performance-based tests such as the six minute walk
S.A.A.N. Bolink et al. / The Knee 22 (2015) 618–623
test (6MWT), timed up-and-go test (TUG) and stair climbing test (SCT) [1]. However, time can be confounded by motivation and instructions and may not fully capture the nature of the impairment that limits physical function. A biomechanical approach could provide more detailed information which potentially can identify OA specific mechanisms causing alterations in movement and compensatory strategies. Various performance-based tests and measurement tools for objective assessment of physical function are available including optical motion capture (MOCAP) systems, force platforms, isometric strength testing, electromyography and ambulant motion sensors [9,10]. In a recent study by Dobson et al. [11], a consensus derived minimal core set of performance-based tests to assess physical function in patients with hip or knee OA has been recommended, including walking short distances, sit-to-stand (STS) transfers and stair negotiation. Furthermore, supplementation with ambulant sensor motion analysis is suggested. Previous work focused on the development of a standardized performance-based test using an ambulant inertial measurement unit (IMU) to assess gait, sit-to-stand transfers and step-up transfers, and demonstrated its potential as a clinical tool for routine functional outcome assessment by comparing patients with knee OA to healthy subjects [3]. This specific method has been adopted and reproduced by other researchers to assess functional outcome following total joint arthroplasty [12–14]. The primary aim of this study was to investigate the responsiveness of IMU derived parameters of physical function during gait, sit–stand transfers and step-up transfers, in a prospectively measured cohort one year after TKA. We hypothesized that for each of the three performance-based tests, significant improvements would be found and that timed parameters and IMU derived parameters would both demonstrate sensitivity to postoperative changes. A second aim was to compare the outcomes of performance-based measures with conventional PROMs. We hypothesized a higher level of responsiveness for PROMs due to instant relief of OA related pain postoperatively (pain dominance [7]) and high expectations of functional improvement (psychological factors [4]). Furthermore we hypothesized that the subjective functional scores of PROMs and objective performance-based measures would correlate moderately as they may capture different dimensions of physical function [1,5]. 2. Materials & methods 2.1. Subjects Patients with end stage knee OA (n = 20, m/f = 7/13; age = 67.4 standard deviation 7.7 years) were randomly recruited from the outpatient clinic if they were listed for a TKA by an orthopaedic surgeon [3]. Ethical approval for testing the patients was obtained. All patients reported activity limitation because of OA, and demonstrated limited knee function on physical examination and their radiographs demonstrated large osteophytes, marked joint space narrowing, severe sclerosis and definite bone contour deformity (Kellgren–Lawrence radiographic OA index scores 3–4) [15]. Physical status of the patients, according to the American Society of Anesthesiologist (ASA)
619
classification system [16] showed that 8/20 patients were ASA class 1 (i.e. healthy person) and 12/20 patients were ASA class 2 (i.e. mild systemic disease). Joint specific comorbidity according to the Charnley classification system [17] showed that 10/20 patients were class A (i.e. single knee with OA); 5/20 patients were class B1 (i.e. bilateral knees with OA) and 5/20 were class B2 (i.e. previous TKA on the contralateral knee). Patients with arthritis of the hip were excluded. Surgery was performed by two different orthopaedic surgeons, each with more than 10 years of experience. All patients underwent a TKA (Vanguard, Biomet) by a medial parapatellar approach. A cruciate-retaining (CR) TKA was used in 13/20 patients whereas a posterior-stabilised (PS) TKA was used in 7/20 patients. Furthermore, in 14/20 TKAs the patella was resurfaced. All patients received physiotherapy for early mobilisation from day one postoperatively, and were discharged at day four postoperatively. Outpatient physical therapy was continued for three months postoperatively and each patient was treated based on his or her individual impairments. 2.2. PROMs The WOMAC [18] and the knee specific KSS [19] were used as patient-reported outcome measures. These PROMs were completed preoperatively and at one year postoperatively. The WOMAC was used because it is reliable, is easy to administer and has been repeatedly used in patients with OA. The WOMAC score is designed to provide information on the patient's perception of pain (five items), stiffness (two items) and physical function (17 items) and each item is scored on a five-point ordered response scale. The score was transformed to a 0–100 score, with 0 representing the best score and 100 representing the worst score [20]. The KSS consists of two domains and each domain is analysed separately: a knee score (0–100) that rates only the knee joint itself and a function score (0–100) that rates the patient's ability to walk and climb stairs. With regard to the knee score, only the three main parameters of pain (50 points), stability (25 points) and range of motion (25 points) are judged and flexion contracture, extension lag and misalignment are dealt with as deductions [19]. With regard to the function score, only walking distance and stair climbing are scored, with deductions for walking aids. Walking distance is expressed in blocks (approximately 100 m) and stair climbing is considered normal if the patient can ascend and descend stairs without holding a railing. The best function score (i.e. 0) is obtained by a patient who can walk an unlimited distance and go up and down stairs normally. 2.3. Physical performance-based tests Physical performance-based tests were undertaken preoperatively and at one year postoperatively, supervised by the same physician. Patients were shown an example of each test but they were not trained for the tests. (1) Gait Subjects walked a 20 m distance in an indoor environment along a straight flat corridor, at their own preferred speed, wearing
Fig. 1. Frontal plane orientation angles during three repetitions of a sit–stand–sit movement.
620
S.A.A.N. Bolink et al. / The Knee 22 (2015) 618–623 40,0
*
5,0
*
Effect size = 1.20 5,0
4,0
30,0
*
6,0
Effect size = 0.84
Effect size = 0.73
4,0
3,0 3,0
20,0
2,0 2,0 10,0
0,0
1,0
Preoperative
Postoperative
1,0 0,0
0,0
Gait
BS
STS
Fig. 2. Time (s) to task completion demonstrating preoperative (dark grey) versus postoperative (light grey) results for gait, STS and BS with and corresponding effect sizes. *p b 0.05.
their own clothes and shoes although high heels were not permitted [21]. None of the subjects used a walking aid. Across the finish line, one last step was allowed to avoid a significant slowdown aiming to reach the marked distance. The exact distance covered (20 m + the last step) was measured and used for further data analysis. Two repetitions were performed and mean values were used for further analysis. (2) Sit-to-stand (STS) Subjects performed three repetitions of STS at preferred speed from a height adjustable chair in a standardized position with hips and knees flexed in a 90° angle, both feet parallel on the floor and arms crossed over the chest [22]. Mean values of three repetitions were used for further analysis. (3) Block step-up (BS) Subjects performed three repetitions for both legs of BS onto a 20 cm high wooden block at preferred speed. Patients first performed three repetitions stepping up with the unaffected leg followed by three repetitions with the affected leg [3,12]. Mean values of three repetitions were used for further analysis.
2.4. Ambulant motion sensor An ambulant inertial measurement unit (IMU; 41 × 63 × 24 mm3; 39 g; MicroStrain® Inertia-Link®) was used containing a magnetometer, gyroscopes (standard deviation 300° s−1) and accelerometers (standard deviation 5 g) along orthogonal axes in frontal, sagittal and transverse planes. Through the IMU's inbuilt fusion algorithms, static and dynamic orientation angles of the device can be derived [23]. Previous studies investigating the agreement between dynamic angle estimation by IMU and MOCAP systems have demonstrated good agreement under laboratory and clinical conditions [23–26]. The IMU was attached onto the skin using a double-sided adhesive tape and positioned at the dorsal side of the pelvis between both posterior superior iliac spines (PSIS) [3,12]. Here, the IMU measures accelerations near the body's centre of mass and dynamic orientation angles of the lower trunk and pelvis (Fig. 1). Hence, it does not allow direct assessment of knee joint function but captures known compensation mechanisms for knee joint dysfunction, such as asymmetrical leg loading and shifting of body weight towards the unaffected side [3,8,12].
derived with accelerometer based HS detection: (1) speed (m/s); (2) cadence (steps/min); (3) step time (s); (4) step length (m); (5) step time SD irregularity (coefficient of variance (cv)): 100% mean ; and (6) step absðleft−rightÞ ; [21,28]. time asymmetry (%):100% ðleftþright Þ 2 Kinematic motion parameters were derived from the IMU's dynamic angle estimation output signals. For gait, the range of motion (RoM; °) in the frontal (medio-lateral) plane (RoMf) and sagittal (antero-posterior) plane (RoMs) was identified for each step separately by computing the HS events to the angle estimation output signals in MATLAB. For STS and BS, kinematic motion parameters represent the mean of three repetitions which were automatically derived by algorithms in MATLAB based on peak detection: time (s) to task completion, RoM (°) in frontal (RoMf) and sagittal (RoMs) planes, and peak accelerations in the antero-posterior (Acc Anteroposterior), medio-lateral (Acc Mediolateral) and vertical (Acc Vertical) directions (m/s2) [29]. 2.6. Statistical analysis Statistical operations were performed with SPSS version 22. The comparison between pre- and postoperative outcomes of the PROMs and performance-based measures was performed with the nonparametric Mann–Whitney U test (p b 0.05). Responsiveness of each outcome measure was assessed by determining the effect size of the difference in means: effect size ¼
absoluteðpreoperative mean−postoperative meanÞ : preoperative standard deviation
Correlations between the subjective functional outcome scores of PROMs and objective physical performance-based parameters were calculated with the non-parametric Spearman's rho (ρ) correlation coefficient. 3. Results In all three performance-based tests, patients demonstrated significant improvement in the time-to-task completion, comparing preoperative with postoperative outcomes for gait (time = 24.27 standard deviation 4.80 s vs. 20.78 standard deviation 4.07 s; effect size = 0.73; p = 0.005), STS (time = 3.19 standard deviation 1.08 s vs. 2.29 standard deviation 0.56 s; effect size = 0.84; p = 0.005) and BS (time = 4.30 standard deviation 1.34 s vs. 2.70 standard deviation 0.42 s; effect size = 1.20; p b 0.001) (Fig. 2). In gait, patients demonstrated significantly higher walking speed (0.85 standard deviation 0.16 m/s vs.
2.5. Data acquisition Via a wireless Bluetooth connection, real-time data from the IMU were stored onto a PC with a sampling frequency of 100 Hz. Data analysis was performed running analysis algorithms in MATLAB® (MathWorks®) version R2009a. For gait, spatiotemporal parameters were derived after heel strike (HS) detection based on the zero crossing method described by Gonzalez et al. [27]. For each walking trial, the first two steps and the last two steps were excluded for further data analysis to avoid the acceleration and deceleration phases as suggested by Bautmans et al. [28]. Based upon the timed task completion for 20 m walking distance and step count, the following spatiotemporal gait parameters were
Table 1 Outcomes of PROMs preoperatively versus postoperatively demonstrating mean values and standard deviation (SD), p-value and effect size. PROMs
WOMAC pain WOMAC stiffness WOMAC function WOMAC total KSS knee KSS function
Preoperative
Postoperative
Mean
SD
Mean
SD
p-Value
Effect size
10.0 3.6 34.4 47.5 50.0 38.5
3.5 2.3 9.6 14.4 15.1 13.5
1.1 1.7 10.0 12.7 21.2 19.0
1.4 1.9 7. 6 9.5 6.3 12.6
b0.001 0.007 b0.001 b0.001 b0.001 b0.001
2.54 0.83 2.54 2.41 1.91 1.45
S.A.A.N. Bolink et al. / The Knee 22 (2015) 618–623 Table 2 Outcome of gait parameters preoperatively versus postoperatively demonstrating mean values and standard deviation (SD), p-value and effect size. Gait
Time (s) Speed (m/s) Cadence (steps/min) Step time (s) Step length (m) Step irregularity (cv) Step asymmetry (%) RoMs (°) RoMf (°)
Preoperative
Postoperative
Mean
SD
Mean
SD
p-Value
Effect size
24.27 0.85 98.05 0.62 0.52 6.13 0.42 5.4 4.9
4.80 0.16 9.81 0.06 0.07 3.34 0.46 1.6 1.8
20.78 0.99 104.73 0.57 0.57 5.99 0.56 6.3 6.6
4.07 0.17 6.38 0.03 0.08 1.84 0.45 1.8 2.4
0.005 0.003 n.s. n.s. 0.026 n.s. n.s. 0.033 0.028
0.73 0.86 0.68 0.68 0.73 0.04 0.30 0.56 0.94
621
Table 4 Outcome of BS (block step) parameters preoperatively versus postoperatively demonstrating mean values and standard deviation (SD), p-value and effect size. BS
Time (s) RoMs (°) RoMf (°) Acc Mediolateral (m/s2) Acc Vertical (m/s2) Acc Anteroposterior (m/s2)
Preoperative
Postoperative
Mean
SD
Mean
SD
p-Value
Effect size
4.30 20.75 13.76 0.60 1.02 0.51
1.34 5.80 3.32 0.15 0.23 0.08
2.70 15.14 11.13 0.48 0.93 0.43
0.42 5.68 3.64 0.12 0.21 0.09
b0.001 b0.001 0.023 0.008 n.s. 0.001
1.20 0.97 0.79 0.73 0.36 1.07
This study has demonstrated that knee OA patients following a TKA perform significantly better at gait, sit-to-stand transfers and block step-
up transfers one year postoperatively. Parameters of time and IMU derived motion parameters both demonstrated significant improvements postoperatively for each three performance-based tests. The magnitude of postoperative improvement was found the largest for parameters of block step-up transfers. Outcomes of performance-based tests were moderately correlated to PROMs preoperatively and postoperatively which may suggest that they measure a different dimension of outcome. Time seems a powerful parameter to measure changes in physical function. In all three tests, time significantly improved postoperatively demonstrating a big effect size (range 0.74–1.20). This justifies the use of the popular, and easy to measure, timed performance-based tests such as the six minute walk test, timed up-and-go test and stair climbing test [3]. However, time represents a composition of physical function but may not capture detailed disease-dependent mechanisms causing functional impairment. This could also explain the correlations between time and the generic self-report measures found for gait and the KSS function score. A person may walk faster with an asymmetric limping pattern, get up faster from a chair and step up faster onto a block with excessive and asymmetric trunk motion compared to a person who demonstrates normal movement. In contrast, power-related magnitudes such as vertical acceleration and movement-related parameters such as range of motion (RoM) of the pelvis can capture specific functional disabilities. In STS, vertical acceleration significantly improved postoperatively (effect size = 0.91). This might indicate that patients generate higher knee extension velocity postoperatively which could be explained by restoration of quadriceps muscle strength or by restoration of preoperative subconscious unloading of the painful leg [8]. In BS, sagittal and frontal RoMs, asymmetry of sagittal RoM, and antero-posterior and mediolateral accelerations significantly decreased after TKA. These findings might reveal specific disabilities and compensation mechanisms suggesting that preoperatively, patients demonstrate excessive mediolateral movement to seek balance during the single leg stance phase, patients use more forward sway of the trunk to create momentum and reduce the loading time on their affected knee as a result of pain and to compensate for decreased quadriceps muscle strength. Furthermore, parameters of BS in particular demonstrated high level of responsiveness suggesting that BS is a more challenging task which captures well the functional disabilities of patients with knee OA preoperatively, and the functional improvement that can be gained following TKA one year postoperatively.
Table 3 Outcome of STS (sit-to-stand) parameters preoperatively versus postoperatively demonstrating mean values and standard deviation (SD), p-value and effect size.
Table 5 Outcome of BS (block step) asymmetry parameters preoperatively versus postoperatively demonstrating mean values and standard deviation (SD), p-value and effect size.
0.99 standard deviation 0.17 m/s; p = 0.003), more step length (0.52 standard deviation 0.07 m vs. 0.57 standard deviation 0.08 m; effect size = 0.73; p = 0.026) and more RoM in the frontal plane (RoMf = 4.9 standard deviation 1.8° vs. 6.6 standard deviation 2.4°; effect size = 0.94; p = 0.028) and the sagittal plane (RoMs = 5.4 standard deviation 1.6° vs. 6.3 standard deviation 1.8°; effect size = 0.56; p = 0.033) one year postoperatively. The gait parameters cadence, step time, step irregularity and step asymmetry demonstrated no significant difference between preoperative and postoperative measures. In STS, higher vertical acceleration was found one year postoperatively (Acc Vertical = 0.54 standard deviation 0.14 (m/s2) vs. 0.67 standard deviation 0.21 (m/s2); effect size = 0.91; p = 0.043). No significant differences were found for RoMs, RoMf, Acc Mediolateral and Acc Anteroposterior (Table 3). In BS, patients demonstrated significantly less RoM in the sagittal plane (RoMs = 20.8 standard deviation 5.8° vs. 15.1 standard deviation 5.6°; effect size = 0.97; p b 0.001) and in the frontal plane (RoMf = 13.8 standard deviation 3.3° vs. 11.1 standard deviation 3.6°; effect size = 0.79; p = 0.023) one year postoperatively (Table 4). In BS, vertical acceleration showed no significant difference between preoperative and postoperative measures but mediolateral and anteroposterior accelerations were both significantly decreased one year postoperatively (Table 4). BS asymmetry scores demonstrated significantly less asymmetry for RoMs (21.2 standard deviation 13.7% vs. 12.9 standard deviation 9.6%; effect size = 0.61; p-value = 0.026) and vertical acceleration (Acc Vertical = 23.8 standard deviation 16.4% vs. 13.9 standard deviation 9.0%; effect size = 0.56; p-value = 0.046). Other BS asymmetry scores were not significantly different however none of the asymmetry scores demonstrated an increase postoperatively (Table 5). Responsiveness to postoperative changes for parameters of time and IMU derived parameters was highest in BS (Fig. 2; Tables 2–4). PROMs demonstrated significant improvement with a high level of responsiveness for postoperative function scores: WOMAC function = 34.4 standard deviation 9.6 vs. 10.0 standard deviation 7.6; effect size = 2.54; and KSS function = 38.5 standard deviation 13.5 vs. 19.0 standard deviation 12.6; effect size = 1.45 resp. (p b 0.001) (Table 1; Fig. 3). In addition, the WOMAC pain score, WOMAC stiffness score and KSS knee score demonstrated significant improvements postoperatively (Table 1). Comparing the preoperative PROM function scores with preoperative performance-based measures demonstrated a significant correlation for WOMAC function score with STS RoMf only (ρ = 0.64). Postoperatively, the WOMAC function score was only significantly correlated to BS asymmetry RoMf (ρ = 0.48). The preoperative KSS function score demonstrated significant correlations with preoperative gait time (ρ = 0.72), gait step length (ρ = 0.74), BS time (ρ = 0.45), BS RoMf (ρ = 0.60) and BS asymmetry RoMf (ρ = 0.62). Postoperatively, the KSS function score was only significantly correlated to parameters of gait: gait time (ρ = 0.54), gait cadence (ρ = 0.50), gait step time (ρ = 0.50), gait irregularity (ρ = 0.62) and gait asymmetry (ρ = 0.58). Only gait time was both pre- and postoperatively correlated with the KSS function score.
4. Discussion
STS
Time (s) RoMs (°) RoMf (°) Acc Mediolateral (m/s2) Acc Vertical (m/s2) Acc Anteroposterior (m/s2)
Preoperative
Postoperative
Mean
SD
Mean
SD
p-Value
Effect size
BS asymmetry
3.19 43.60 6.49 0.21 0.54 0.73
1.08 10.75 2.94 0.07 0.14 0.18
2.29 45.71 5.33 0.22 0.67 0.80
0.56 12.65 2.23 0.07 0.21 0.19
0.005 n.s. n.s. n.s. 0.043 n.s.
0.84 0.20 0.39 0.00 0.91 0.37
Time (%) RoMs (%) RoMf (%) Acc Mediolateral (%) Acc Vertical (%) Acc Anteroposterior (%)
Preoperative
Postoperative
Mean
SD
Mean
SD
p-Value
Effect size
20.1 21.2 17.4 19.5 23.8 22.1
19.5 13.7 13.0 15.9 16.4 16.5
15.0 12.9 17.3 19.1 13.9 14.5
18.2 9.6 10.2 13.4 9.0 11.0
n.s. 0.026 n.s. n.s. 0.046 n.s.
0.26 0.61 0.01 0.03 0.56 0.46
622
S.A.A.N. Bolink et al. / The Knee 22 (2015) 618–623
50,0
60,0
*
Effect size = 2.54 40,0
*
Effect size = 1.45
50,0 40,0
30,0
30,0 20,0 20,0
10,0 0,0
10,0 preoperative
postoperative
0,0 KSS function
WOMAC function
Fig. 3. Outcome of WOMAC function score and KSS function score preoperatively (dark grey) versus postoperatively (light grey) and corresponding effect sizes. *p b 0.001.
PROMs are most frequently used to evaluate the functional outcome after TKA. However, PROMs reflect a patient's perception of function and there is increasing evidence to suggest that reported pain instead of actual physical function is the main determinant of PROMs [30]. Solely using PROMs for outcome assessment after TKA seems to overestimate the actual short-term and long-term changes in physical function due to instant pain relief postoperatively and high expectations [4,7,30]. These inherent problems with PROMs have made objectively assessed instruments of functional outcome more appealing to clinicians. Previous studies have shown that during the first month in postoperative recovery, patients perform worse at various performance-based tests compared to their preoperative levels but in contrast, PROMs demonstrate a significant improvement [1,5,31]. Moreover, Mizner et al. [31] demonstrated that performance-based measures yield greater responsiveness during the acute postoperative stages (i.e. first month) with a negative effect size (i.e. worse performance), whereas the long-term (i.e. one year) responsiveness with a positive effect size is greater for PROMs. The findings of this study are consistent with those of previous reports demonstrating a smaller effect size for performance-based measures compared to PROMs at one year postoperatively. In addition, PROMs and some of the performance-based measures were moderately correlated preoperatively and postoperatively. This suggests that both measures capture physical function with some redundancy but they may also capture other dimensions of physical function and could potentially supplement each other. Our study results demonstrate that the WOMAC function score was only moderately correlated to RoMf during STS preoperatively, and to the asymmetry of RoMf during BS postoperatively. The KSS function score demonstrated more significant correlations to performancebased measures pre- and postoperatively, especially for parameters of gait. However, only gait time was both pre- and postoperatively significantly correlated to the KSS function score. The moderate correlations (Spearman's rho range 0.45–0.74) between the PROMs and performance-based measures may highlight the inherent link yet relative independence of the two measurement tools. PROMs are generic measures of outcome and include items across a broad range of health aspects. Therefore, PROMs may lack the level of detail in terms of relevance to any specific dimension measured with performance-based measures which perhaps accounts for the lower responsiveness and weaker correlations. The results of this study suggest that patients seem to score their physical performance best by reporting about their walking abilities, which may be best captured by the KSS function score. The study has some limitations that merit attention when interpreting the results. The cohort of patients undergoing a primary TKA in this study showed good outcomes of PROMs in comparison to previous findings [5,32], and there is a potential that the responsiveness of the outcome measures is slightly overstated compared to other populations. This might be due to the rather good physical status of this study's cohort as 8/20 patients were ASA class 1 and 10/20 patients had no other joint affected by OA (Charnley class A). Preoperative ASA
physical status score and Charnley classification have demonstrated to influence postoperative functional outcome following TKA [16,17]. Postoperative physical therapy was not standardized which could account for individual differences. Furthermore, as the population does not represent a consecutive case series, selection bias could be a confounder. Other limitations include the small sample size and the lack of more follow-up moments. Nonetheless, the prospective nature of the study, the use of a standardized performance-based test with novel IMU based outcome measures and the comparison with conventional patient-reported outcome measures are a strength and provide more insight into postoperative TKA recovery for comprehensive assessment of functional outcome. 5. Conclusion PROMs and physical performance-based outcome measures are both responsive to changes one year after TKA, but the moderate correlations may suggest that they measure a different dimension of outcome. Performance-based measures are necessary to objectively assess the patients' actual physical function that is not captured by PROMs and to fully characterize the changes in physical function of patients after TKA, particularly in younger or high demand patients where ceiling effects may be a problem. A physically challenging task, such as stepping onto a block, seems to capture the functional disabilities of patients with knee OA and the functional improvement after TKA better than a less challenging task, such as walking a short distance. To capture the variety of a patient's functional disability, we suggest that a routine clinical performance-based test should equally consist of a variety of daily life activities. Moreover, we suggest that a performance-based test should not only represent time-related measures but comprise a biomechanical approach capturing the underlying mechanisms causing functional impairment, to guide more targeted rehabilitation and to improve overall functional outcome. Conflict of interest statement None. References [1] Stevens-Lapsley JE, Schenkman ML, Dayton MR. Comparison of self-reported knee injury and osteoarthritis outcome score to performance measures in patients after total knee arthroplasty. PMR 2011;3(6):541–9 [quiz 549]. [2] Nilsdotter AK, Toksvig-Larsen S, Roos EM. Knee arthroplasty: are patients' expectations fulfilled? A prospective study of pain and function in 102 patients with 5year follow-up. Acta Orthop 2009;80(1):55–61. [3] Bolink SA, et al. Inertial sensor motion analysis of gait, sit–stand transfers and stepup transfers: differentiating knee patients from healthy controls. Physiol Meas 2012; 33(11):1947–58. [4] Vissers MM, et al. Psychological factors affecting the outcome of total hip and knee arthroplasty: a systematic review. Semin Arthritis Rheum 2012;41(4):576–88. [5] Senden R, et al. The importance to including objective functional outcomes in the clinical follow up of total knee arthroplasty patients. Knee 2011;18(5):306–11.
S.A.A.N. Bolink et al. / The Knee 22 (2015) 618–623 [6] Ornetti P, et al. Gait analysis as a quantifiable outcome measure in hip or knee osteoarthritis: a systematic review. Joint Bone Spine 2010;77(5):421–5. [7] Terwee CB, et al. Self-reported physical functioning was more influenced by pain than performance-based physical functioning in knee-osteoarthritis patients. J Clin Epidemiol 2006;59(7):724–31. [8] Boonstra MC, De Waal Malefijt MC, Verdonschot N. How to quantify knee function after total knee arthroplasty? Knee 2008;15(5):390–5. [9] Dobson F, et al. Measurement properties of performance-based measures to assess physical function in hip and knee osteoarthritis: a systematic review. Osteoarthritis Cartilage 2012;20(12):1548–62. [10] van den Dikkenberg N, et al. Measuring functional abilities of patients with knee problems: rationale and construction of the DynaPort knee test. Knee Surg Sports Traumatol Arthrosc 2002;10(4):204–12. [11] Dobson F, et al. OARSI recommended performance-based tests to assess physical function in people diagnosed with hip or knee osteoarthritis. Osteoarthritis Cartilage 2013;21(8):1042–52. [12] Hjorth MH, et al. Block-step asymmetry 5 years after large-head metal-on-metal total hip arthroplasty is related to lower muscle mass and leg power on the implant side. Clin Biomech (Bristol, Avon) 2014;29(6):684–90. [13] Wylde V, et al. Assessing function in patients undergoing joint replacement: a study protocol for a cohort study. BMC Musculoskelet Disord 2012;13:220. [14] Wylde V, et al. Does measuring the range of motion of the hip and knee add to the assessment of disability in people undergoing joint replacement? Orthop Traumatol Surg Res 2014;100(2):183–6. [15] Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957;16(4):494–502. [16] Hooper GJ, et al. The relationship between the American Society Of Anesthesiologists physical rating and outcome following total hip and knee arthroplasty: an analysis of the New Zealand Joint Registry. J Bone Joint Surg Am 2012;94(12):1065–70. [17] Bjorgul K, Novicoff WM, Saleh KJ. Evaluating comorbidities in total hip and knee arthroplasty: available instruments. J Orthop Traumatol 2010;11(4):203–9. [18] Bellamy N, et al. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol 1988;15(12):1833–40.
623
[19] Insall JN, et al. Rationale of the Knee Society clinical rating system. Clin Orthop Relat Res 1989;248:13–4. [20] Unnanuntana A, et al. Performance-based tests and self-reported questionnaires provide distinct information for the preoperative evaluation of total hip arthroplasty patients. J Arthroplasty 2012;27(5):770–5 [e1]. [21] Senden R, et al. Acceleration-based gait test for healthy subjects: reliability and reference data. Gait Posture 2009;30(2):192–6. [22] Janssen WG, Bussmann HB, Stam HJ. Determinants of the sit-to-stand movement: a review. Phys Ther 2002;82(9):866–79. [23] Lebel K, et al. Inertial measures of motion for clinical biomechanics: comparative assessment of accuracy under controlled conditions — effect of velocity. PLoS One 2013;8(11):e79945. [24] Takeda R, et al. Gait posture estimation using wearable acceleration and gyro sensors. J Biomech 2009;42(15):2486–94. [25] Seel T, Raisch J, Schauer T. IMU-based joint angle measurement for gait analysis. Sensors (Basel) 2014;14(4):6891–909. [26] Wong WY, Wong MS. Trunk posture monitoring with inertial sensors. Eur Spine J 2008;17(5):743–53. [27] Gonzalez RC, et al. Real-time gait event detection for normal subjects from lower trunk accelerations. Gait Posture 2010;31(3):322–5. [28] Bautmans I, et al. Reliability and clinical correlates of 3D-accelerometry based gait analysis outcomes according to age and fall-risk. Gait Posture 2011;33(3):366–72. [29] Boonstra MC, et al. Sit-to-stand movement as a performance-based measure for patients with total knee arthroplasty. Phys Ther 2010;90(2):149–56. [30] Hossain FS, et al. A performance based patient outcome score for active patients following total knee arthroplasty. Osteoarthritis Cartilage 2013;21(1):51–9. [31] Mizner RL, et al. Measuring functional improvement after total knee arthroplasty requires both performance-based and patient-report assessments: a longitudinal analysis of outcomes. J Arthroplasty 2011;26(5):728–37. [32] Khanna G, et al. Comparison of patient-reported and clinician-assessed outcomes following total knee arthroplasty. J Bone Joint Surg Am 2011;93(20):e117 [(1)–(7)].
Contents lists available at ScienceDirect
The Knee
Patient-reported outcome measures versus inertial performance-based outcome measures: A prospective study in patients undergoing primary total knee arthroplasty S.A.A.N. Bolink ⁎, B. Grimm, I.C. Heyligers AHORSE Foundation, Dept. of Orthopaedics, Atrium Medical Center Heerlen, The Netherlands
a r t i c l e
i n f o
Article history: Received 10 December 2014 Received in revised form 11 March 2015 Accepted 15 April 2015 Keywords: Total knee arthroplasty Outcome assessment Performance-based test PROMs Inertial measurement unit
a b s t r a c t Background: Outcome assessment of total knee arthroplasty (TKA) by subjective patient reported outcome measures (PROMs) may not fully capture the functional (dis-)abilities of relevance. Objective performance-based outcome measures could provide distinct information. An ambulant inertial measurement unit (IMU) allows kinematic assessment of physical performance and could potentially be used for routine follow-up. Aim: To investigate the responsiveness of IMU measures in patients following TKA and compare outcomes with conventional PROMs. Methods: Patients with end stage knee OA (n = 20, m/f = 7/13; age = 67.4 standard deviation 7.7 years) were measured preoperatively and one year postoperatively. IMU measures were derived during gait, sit–stand transfers and block step-up transfers. PROMs were assessed by using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and Knee Society Score (KSS). Responsiveness was calculated by the effect size, correlations were calculated with Spearman's rho correlation coefficient. Results: One year after TKA, patients performed significantly better at gait, sit-to-stand transfers and block stepup transfers. Measures of time and kinematic IMU measures demonstrated significant improvements postoperatively for each performance-based test. The largest improvement was found in block step-up transfers (effect size = 0.56–1.20). WOMAC function score and KSS function score demonstrated moderate correlations (Spearman's rho = 0.45–0.74) with some of the physical performance-based measures pre- and postoperatively. Conclusion: To characterize the changes in physical function after TKA, PROMs could be supplemented by performance-based measures, assessing function during different activities and allowing kinematic characterization with an ambulant IMU. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Total knee arthroplasty (TKA) has developed into a successful intervention for patients with advanced knee osteoarthritis (OA). The incidence of knee OA and the number of TKAs are expected to increase substantially in the future [1] with a bigger proportion of relatively young patients who are more active and more demanding [2]. This challenges the outcome assessment methods we use, which equally need to adapt and evolve [3]. Traditionally, outcome assessment after TKA includes pain and function and it is important to have separate measures of both domains. Pain is often considered the most important aspect of outcome however it is a subjective measure and influenced by socioeconomic and psychosocial factors [4]. Function is becoming more important, especially for the new generation of physically more active patients, and can be assessed by either patient-reported measures ⁎ Corresponding author at: Dept. of Orthopaedics, Atrium Medical Center Heerlen, Henri Dunantstraat 5, 6419 PC Heerlen, The Netherlands. Tel.: +31 655341492. E-mail address: [email protected] (S.A.A.N. Bolink).
http://dx.doi.org/10.1016/j.knee.2015.04.002 0968-0160/© 2015 Elsevier B.V. All rights reserved.
(e.g. questionnaires) and performance-based measures (e.g. gait analysis) [5,6]. Popular patient-reported outcome measures (PROMs), such as the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and Knee Society Score (KSS), combine measures of pain and function and are convenient to use in longitudinal follow-up. However, PROMs are subjective measures, suffer from a ceiling effect and their pain dominance masks the functional changes [7]. Therefore, outcome assessment by PROMs only may not fully capture the functional (dis-)abilities of relevance. Performance-based outcome measures on the other hand tend to be more objective and measure what a patient actually can do [8]. As performance-based measures have a low correlation with PROMs, they may capture other dimensions of physical function and could provide distinct and supplementary information [5]. Thus, PROMs and performance-based tests are both essential components of comprehensive functional outcome assessment for longitudinal follow-up. Unfortunately, performance-based measures struggle with feasibility issues as they are mostly time-consuming and require dedicated laboratories. This justifies the frequently used, and easy to measure, timed performance-based tests such as the six minute walk
S.A.A.N. Bolink et al. / The Knee 22 (2015) 618–623
test (6MWT), timed up-and-go test (TUG) and stair climbing test (SCT) [1]. However, time can be confounded by motivation and instructions and may not fully capture the nature of the impairment that limits physical function. A biomechanical approach could provide more detailed information which potentially can identify OA specific mechanisms causing alterations in movement and compensatory strategies. Various performance-based tests and measurement tools for objective assessment of physical function are available including optical motion capture (MOCAP) systems, force platforms, isometric strength testing, electromyography and ambulant motion sensors [9,10]. In a recent study by Dobson et al. [11], a consensus derived minimal core set of performance-based tests to assess physical function in patients with hip or knee OA has been recommended, including walking short distances, sit-to-stand (STS) transfers and stair negotiation. Furthermore, supplementation with ambulant sensor motion analysis is suggested. Previous work focused on the development of a standardized performance-based test using an ambulant inertial measurement unit (IMU) to assess gait, sit-to-stand transfers and step-up transfers, and demonstrated its potential as a clinical tool for routine functional outcome assessment by comparing patients with knee OA to healthy subjects [3]. This specific method has been adopted and reproduced by other researchers to assess functional outcome following total joint arthroplasty [12–14]. The primary aim of this study was to investigate the responsiveness of IMU derived parameters of physical function during gait, sit–stand transfers and step-up transfers, in a prospectively measured cohort one year after TKA. We hypothesized that for each of the three performance-based tests, significant improvements would be found and that timed parameters and IMU derived parameters would both demonstrate sensitivity to postoperative changes. A second aim was to compare the outcomes of performance-based measures with conventional PROMs. We hypothesized a higher level of responsiveness for PROMs due to instant relief of OA related pain postoperatively (pain dominance [7]) and high expectations of functional improvement (psychological factors [4]). Furthermore we hypothesized that the subjective functional scores of PROMs and objective performance-based measures would correlate moderately as they may capture different dimensions of physical function [1,5]. 2. Materials & methods 2.1. Subjects Patients with end stage knee OA (n = 20, m/f = 7/13; age = 67.4 standard deviation 7.7 years) were randomly recruited from the outpatient clinic if they were listed for a TKA by an orthopaedic surgeon [3]. Ethical approval for testing the patients was obtained. All patients reported activity limitation because of OA, and demonstrated limited knee function on physical examination and their radiographs demonstrated large osteophytes, marked joint space narrowing, severe sclerosis and definite bone contour deformity (Kellgren–Lawrence radiographic OA index scores 3–4) [15]. Physical status of the patients, according to the American Society of Anesthesiologist (ASA)
619
classification system [16] showed that 8/20 patients were ASA class 1 (i.e. healthy person) and 12/20 patients were ASA class 2 (i.e. mild systemic disease). Joint specific comorbidity according to the Charnley classification system [17] showed that 10/20 patients were class A (i.e. single knee with OA); 5/20 patients were class B1 (i.e. bilateral knees with OA) and 5/20 were class B2 (i.e. previous TKA on the contralateral knee). Patients with arthritis of the hip were excluded. Surgery was performed by two different orthopaedic surgeons, each with more than 10 years of experience. All patients underwent a TKA (Vanguard, Biomet) by a medial parapatellar approach. A cruciate-retaining (CR) TKA was used in 13/20 patients whereas a posterior-stabilised (PS) TKA was used in 7/20 patients. Furthermore, in 14/20 TKAs the patella was resurfaced. All patients received physiotherapy for early mobilisation from day one postoperatively, and were discharged at day four postoperatively. Outpatient physical therapy was continued for three months postoperatively and each patient was treated based on his or her individual impairments. 2.2. PROMs The WOMAC [18] and the knee specific KSS [19] were used as patient-reported outcome measures. These PROMs were completed preoperatively and at one year postoperatively. The WOMAC was used because it is reliable, is easy to administer and has been repeatedly used in patients with OA. The WOMAC score is designed to provide information on the patient's perception of pain (five items), stiffness (two items) and physical function (17 items) and each item is scored on a five-point ordered response scale. The score was transformed to a 0–100 score, with 0 representing the best score and 100 representing the worst score [20]. The KSS consists of two domains and each domain is analysed separately: a knee score (0–100) that rates only the knee joint itself and a function score (0–100) that rates the patient's ability to walk and climb stairs. With regard to the knee score, only the three main parameters of pain (50 points), stability (25 points) and range of motion (25 points) are judged and flexion contracture, extension lag and misalignment are dealt with as deductions [19]. With regard to the function score, only walking distance and stair climbing are scored, with deductions for walking aids. Walking distance is expressed in blocks (approximately 100 m) and stair climbing is considered normal if the patient can ascend and descend stairs without holding a railing. The best function score (i.e. 0) is obtained by a patient who can walk an unlimited distance and go up and down stairs normally. 2.3. Physical performance-based tests Physical performance-based tests were undertaken preoperatively and at one year postoperatively, supervised by the same physician. Patients were shown an example of each test but they were not trained for the tests. (1) Gait Subjects walked a 20 m distance in an indoor environment along a straight flat corridor, at their own preferred speed, wearing
Fig. 1. Frontal plane orientation angles during three repetitions of a sit–stand–sit movement.
620
S.A.A.N. Bolink et al. / The Knee 22 (2015) 618–623 40,0
*
5,0
*
Effect size = 1.20 5,0
4,0
30,0
*
6,0
Effect size = 0.84
Effect size = 0.73
4,0
3,0 3,0
20,0
2,0 2,0 10,0
0,0
1,0
Preoperative
Postoperative
1,0 0,0
0,0
Gait
BS
STS
Fig. 2. Time (s) to task completion demonstrating preoperative (dark grey) versus postoperative (light grey) results for gait, STS and BS with and corresponding effect sizes. *p b 0.05.
their own clothes and shoes although high heels were not permitted [21]. None of the subjects used a walking aid. Across the finish line, one last step was allowed to avoid a significant slowdown aiming to reach the marked distance. The exact distance covered (20 m + the last step) was measured and used for further data analysis. Two repetitions were performed and mean values were used for further analysis. (2) Sit-to-stand (STS) Subjects performed three repetitions of STS at preferred speed from a height adjustable chair in a standardized position with hips and knees flexed in a 90° angle, both feet parallel on the floor and arms crossed over the chest [22]. Mean values of three repetitions were used for further analysis. (3) Block step-up (BS) Subjects performed three repetitions for both legs of BS onto a 20 cm high wooden block at preferred speed. Patients first performed three repetitions stepping up with the unaffected leg followed by three repetitions with the affected leg [3,12]. Mean values of three repetitions were used for further analysis.
2.4. Ambulant motion sensor An ambulant inertial measurement unit (IMU; 41 × 63 × 24 mm3; 39 g; MicroStrain® Inertia-Link®) was used containing a magnetometer, gyroscopes (standard deviation 300° s−1) and accelerometers (standard deviation 5 g) along orthogonal axes in frontal, sagittal and transverse planes. Through the IMU's inbuilt fusion algorithms, static and dynamic orientation angles of the device can be derived [23]. Previous studies investigating the agreement between dynamic angle estimation by IMU and MOCAP systems have demonstrated good agreement under laboratory and clinical conditions [23–26]. The IMU was attached onto the skin using a double-sided adhesive tape and positioned at the dorsal side of the pelvis between both posterior superior iliac spines (PSIS) [3,12]. Here, the IMU measures accelerations near the body's centre of mass and dynamic orientation angles of the lower trunk and pelvis (Fig. 1). Hence, it does not allow direct assessment of knee joint function but captures known compensation mechanisms for knee joint dysfunction, such as asymmetrical leg loading and shifting of body weight towards the unaffected side [3,8,12].
derived with accelerometer based HS detection: (1) speed (m/s); (2) cadence (steps/min); (3) step time (s); (4) step length (m); (5) step time SD irregularity (coefficient of variance (cv)): 100% mean ; and (6) step absðleft−rightÞ ; [21,28]. time asymmetry (%):100% ðleftþright Þ 2 Kinematic motion parameters were derived from the IMU's dynamic angle estimation output signals. For gait, the range of motion (RoM; °) in the frontal (medio-lateral) plane (RoMf) and sagittal (antero-posterior) plane (RoMs) was identified for each step separately by computing the HS events to the angle estimation output signals in MATLAB. For STS and BS, kinematic motion parameters represent the mean of three repetitions which were automatically derived by algorithms in MATLAB based on peak detection: time (s) to task completion, RoM (°) in frontal (RoMf) and sagittal (RoMs) planes, and peak accelerations in the antero-posterior (Acc Anteroposterior), medio-lateral (Acc Mediolateral) and vertical (Acc Vertical) directions (m/s2) [29]. 2.6. Statistical analysis Statistical operations were performed with SPSS version 22. The comparison between pre- and postoperative outcomes of the PROMs and performance-based measures was performed with the nonparametric Mann–Whitney U test (p b 0.05). Responsiveness of each outcome measure was assessed by determining the effect size of the difference in means: effect size ¼
absoluteðpreoperative mean−postoperative meanÞ : preoperative standard deviation
Correlations between the subjective functional outcome scores of PROMs and objective physical performance-based parameters were calculated with the non-parametric Spearman's rho (ρ) correlation coefficient. 3. Results In all three performance-based tests, patients demonstrated significant improvement in the time-to-task completion, comparing preoperative with postoperative outcomes for gait (time = 24.27 standard deviation 4.80 s vs. 20.78 standard deviation 4.07 s; effect size = 0.73; p = 0.005), STS (time = 3.19 standard deviation 1.08 s vs. 2.29 standard deviation 0.56 s; effect size = 0.84; p = 0.005) and BS (time = 4.30 standard deviation 1.34 s vs. 2.70 standard deviation 0.42 s; effect size = 1.20; p b 0.001) (Fig. 2). In gait, patients demonstrated significantly higher walking speed (0.85 standard deviation 0.16 m/s vs.
2.5. Data acquisition Via a wireless Bluetooth connection, real-time data from the IMU were stored onto a PC with a sampling frequency of 100 Hz. Data analysis was performed running analysis algorithms in MATLAB® (MathWorks®) version R2009a. For gait, spatiotemporal parameters were derived after heel strike (HS) detection based on the zero crossing method described by Gonzalez et al. [27]. For each walking trial, the first two steps and the last two steps were excluded for further data analysis to avoid the acceleration and deceleration phases as suggested by Bautmans et al. [28]. Based upon the timed task completion for 20 m walking distance and step count, the following spatiotemporal gait parameters were
Table 1 Outcomes of PROMs preoperatively versus postoperatively demonstrating mean values and standard deviation (SD), p-value and effect size. PROMs
WOMAC pain WOMAC stiffness WOMAC function WOMAC total KSS knee KSS function
Preoperative
Postoperative
Mean
SD
Mean
SD
p-Value
Effect size
10.0 3.6 34.4 47.5 50.0 38.5
3.5 2.3 9.6 14.4 15.1 13.5
1.1 1.7 10.0 12.7 21.2 19.0
1.4 1.9 7. 6 9.5 6.3 12.6
b0.001 0.007 b0.001 b0.001 b0.001 b0.001
2.54 0.83 2.54 2.41 1.91 1.45
S.A.A.N. Bolink et al. / The Knee 22 (2015) 618–623 Table 2 Outcome of gait parameters preoperatively versus postoperatively demonstrating mean values and standard deviation (SD), p-value and effect size. Gait
Time (s) Speed (m/s) Cadence (steps/min) Step time (s) Step length (m) Step irregularity (cv) Step asymmetry (%) RoMs (°) RoMf (°)
Preoperative
Postoperative
Mean
SD
Mean
SD
p-Value
Effect size
24.27 0.85 98.05 0.62 0.52 6.13 0.42 5.4 4.9
4.80 0.16 9.81 0.06 0.07 3.34 0.46 1.6 1.8
20.78 0.99 104.73 0.57 0.57 5.99 0.56 6.3 6.6
4.07 0.17 6.38 0.03 0.08 1.84 0.45 1.8 2.4
0.005 0.003 n.s. n.s. 0.026 n.s. n.s. 0.033 0.028
0.73 0.86 0.68 0.68 0.73 0.04 0.30 0.56 0.94
621
Table 4 Outcome of BS (block step) parameters preoperatively versus postoperatively demonstrating mean values and standard deviation (SD), p-value and effect size. BS
Time (s) RoMs (°) RoMf (°) Acc Mediolateral (m/s2) Acc Vertical (m/s2) Acc Anteroposterior (m/s2)
Preoperative
Postoperative
Mean
SD
Mean
SD
p-Value
Effect size
4.30 20.75 13.76 0.60 1.02 0.51
1.34 5.80 3.32 0.15 0.23 0.08
2.70 15.14 11.13 0.48 0.93 0.43
0.42 5.68 3.64 0.12 0.21 0.09
b0.001 b0.001 0.023 0.008 n.s. 0.001
1.20 0.97 0.79 0.73 0.36 1.07
This study has demonstrated that knee OA patients following a TKA perform significantly better at gait, sit-to-stand transfers and block step-
up transfers one year postoperatively. Parameters of time and IMU derived motion parameters both demonstrated significant improvements postoperatively for each three performance-based tests. The magnitude of postoperative improvement was found the largest for parameters of block step-up transfers. Outcomes of performance-based tests were moderately correlated to PROMs preoperatively and postoperatively which may suggest that they measure a different dimension of outcome. Time seems a powerful parameter to measure changes in physical function. In all three tests, time significantly improved postoperatively demonstrating a big effect size (range 0.74–1.20). This justifies the use of the popular, and easy to measure, timed performance-based tests such as the six minute walk test, timed up-and-go test and stair climbing test [3]. However, time represents a composition of physical function but may not capture detailed disease-dependent mechanisms causing functional impairment. This could also explain the correlations between time and the generic self-report measures found for gait and the KSS function score. A person may walk faster with an asymmetric limping pattern, get up faster from a chair and step up faster onto a block with excessive and asymmetric trunk motion compared to a person who demonstrates normal movement. In contrast, power-related magnitudes such as vertical acceleration and movement-related parameters such as range of motion (RoM) of the pelvis can capture specific functional disabilities. In STS, vertical acceleration significantly improved postoperatively (effect size = 0.91). This might indicate that patients generate higher knee extension velocity postoperatively which could be explained by restoration of quadriceps muscle strength or by restoration of preoperative subconscious unloading of the painful leg [8]. In BS, sagittal and frontal RoMs, asymmetry of sagittal RoM, and antero-posterior and mediolateral accelerations significantly decreased after TKA. These findings might reveal specific disabilities and compensation mechanisms suggesting that preoperatively, patients demonstrate excessive mediolateral movement to seek balance during the single leg stance phase, patients use more forward sway of the trunk to create momentum and reduce the loading time on their affected knee as a result of pain and to compensate for decreased quadriceps muscle strength. Furthermore, parameters of BS in particular demonstrated high level of responsiveness suggesting that BS is a more challenging task which captures well the functional disabilities of patients with knee OA preoperatively, and the functional improvement that can be gained following TKA one year postoperatively.
Table 3 Outcome of STS (sit-to-stand) parameters preoperatively versus postoperatively demonstrating mean values and standard deviation (SD), p-value and effect size.
Table 5 Outcome of BS (block step) asymmetry parameters preoperatively versus postoperatively demonstrating mean values and standard deviation (SD), p-value and effect size.
0.99 standard deviation 0.17 m/s; p = 0.003), more step length (0.52 standard deviation 0.07 m vs. 0.57 standard deviation 0.08 m; effect size = 0.73; p = 0.026) and more RoM in the frontal plane (RoMf = 4.9 standard deviation 1.8° vs. 6.6 standard deviation 2.4°; effect size = 0.94; p = 0.028) and the sagittal plane (RoMs = 5.4 standard deviation 1.6° vs. 6.3 standard deviation 1.8°; effect size = 0.56; p = 0.033) one year postoperatively. The gait parameters cadence, step time, step irregularity and step asymmetry demonstrated no significant difference between preoperative and postoperative measures. In STS, higher vertical acceleration was found one year postoperatively (Acc Vertical = 0.54 standard deviation 0.14 (m/s2) vs. 0.67 standard deviation 0.21 (m/s2); effect size = 0.91; p = 0.043). No significant differences were found for RoMs, RoMf, Acc Mediolateral and Acc Anteroposterior (Table 3). In BS, patients demonstrated significantly less RoM in the sagittal plane (RoMs = 20.8 standard deviation 5.8° vs. 15.1 standard deviation 5.6°; effect size = 0.97; p b 0.001) and in the frontal plane (RoMf = 13.8 standard deviation 3.3° vs. 11.1 standard deviation 3.6°; effect size = 0.79; p = 0.023) one year postoperatively (Table 4). In BS, vertical acceleration showed no significant difference between preoperative and postoperative measures but mediolateral and anteroposterior accelerations were both significantly decreased one year postoperatively (Table 4). BS asymmetry scores demonstrated significantly less asymmetry for RoMs (21.2 standard deviation 13.7% vs. 12.9 standard deviation 9.6%; effect size = 0.61; p-value = 0.026) and vertical acceleration (Acc Vertical = 23.8 standard deviation 16.4% vs. 13.9 standard deviation 9.0%; effect size = 0.56; p-value = 0.046). Other BS asymmetry scores were not significantly different however none of the asymmetry scores demonstrated an increase postoperatively (Table 5). Responsiveness to postoperative changes for parameters of time and IMU derived parameters was highest in BS (Fig. 2; Tables 2–4). PROMs demonstrated significant improvement with a high level of responsiveness for postoperative function scores: WOMAC function = 34.4 standard deviation 9.6 vs. 10.0 standard deviation 7.6; effect size = 2.54; and KSS function = 38.5 standard deviation 13.5 vs. 19.0 standard deviation 12.6; effect size = 1.45 resp. (p b 0.001) (Table 1; Fig. 3). In addition, the WOMAC pain score, WOMAC stiffness score and KSS knee score demonstrated significant improvements postoperatively (Table 1). Comparing the preoperative PROM function scores with preoperative performance-based measures demonstrated a significant correlation for WOMAC function score with STS RoMf only (ρ = 0.64). Postoperatively, the WOMAC function score was only significantly correlated to BS asymmetry RoMf (ρ = 0.48). The preoperative KSS function score demonstrated significant correlations with preoperative gait time (ρ = 0.72), gait step length (ρ = 0.74), BS time (ρ = 0.45), BS RoMf (ρ = 0.60) and BS asymmetry RoMf (ρ = 0.62). Postoperatively, the KSS function score was only significantly correlated to parameters of gait: gait time (ρ = 0.54), gait cadence (ρ = 0.50), gait step time (ρ = 0.50), gait irregularity (ρ = 0.62) and gait asymmetry (ρ = 0.58). Only gait time was both pre- and postoperatively correlated with the KSS function score.
4. Discussion
STS
Time (s) RoMs (°) RoMf (°) Acc Mediolateral (m/s2) Acc Vertical (m/s2) Acc Anteroposterior (m/s2)
Preoperative
Postoperative
Mean
SD
Mean
SD
p-Value
Effect size
BS asymmetry
3.19 43.60 6.49 0.21 0.54 0.73
1.08 10.75 2.94 0.07 0.14 0.18
2.29 45.71 5.33 0.22 0.67 0.80
0.56 12.65 2.23 0.07 0.21 0.19
0.005 n.s. n.s. n.s. 0.043 n.s.
0.84 0.20 0.39 0.00 0.91 0.37
Time (%) RoMs (%) RoMf (%) Acc Mediolateral (%) Acc Vertical (%) Acc Anteroposterior (%)
Preoperative
Postoperative
Mean
SD
Mean
SD
p-Value
Effect size
20.1 21.2 17.4 19.5 23.8 22.1
19.5 13.7 13.0 15.9 16.4 16.5
15.0 12.9 17.3 19.1 13.9 14.5
18.2 9.6 10.2 13.4 9.0 11.0
n.s. 0.026 n.s. n.s. 0.046 n.s.
0.26 0.61 0.01 0.03 0.56 0.46
622
S.A.A.N. Bolink et al. / The Knee 22 (2015) 618–623
50,0
60,0
*
Effect size = 2.54 40,0
*
Effect size = 1.45
50,0 40,0
30,0
30,0 20,0 20,0
10,0 0,0
10,0 preoperative
postoperative
0,0 KSS function
WOMAC function
Fig. 3. Outcome of WOMAC function score and KSS function score preoperatively (dark grey) versus postoperatively (light grey) and corresponding effect sizes. *p b 0.001.
PROMs are most frequently used to evaluate the functional outcome after TKA. However, PROMs reflect a patient's perception of function and there is increasing evidence to suggest that reported pain instead of actual physical function is the main determinant of PROMs [30]. Solely using PROMs for outcome assessment after TKA seems to overestimate the actual short-term and long-term changes in physical function due to instant pain relief postoperatively and high expectations [4,7,30]. These inherent problems with PROMs have made objectively assessed instruments of functional outcome more appealing to clinicians. Previous studies have shown that during the first month in postoperative recovery, patients perform worse at various performance-based tests compared to their preoperative levels but in contrast, PROMs demonstrate a significant improvement [1,5,31]. Moreover, Mizner et al. [31] demonstrated that performance-based measures yield greater responsiveness during the acute postoperative stages (i.e. first month) with a negative effect size (i.e. worse performance), whereas the long-term (i.e. one year) responsiveness with a positive effect size is greater for PROMs. The findings of this study are consistent with those of previous reports demonstrating a smaller effect size for performance-based measures compared to PROMs at one year postoperatively. In addition, PROMs and some of the performance-based measures were moderately correlated preoperatively and postoperatively. This suggests that both measures capture physical function with some redundancy but they may also capture other dimensions of physical function and could potentially supplement each other. Our study results demonstrate that the WOMAC function score was only moderately correlated to RoMf during STS preoperatively, and to the asymmetry of RoMf during BS postoperatively. The KSS function score demonstrated more significant correlations to performancebased measures pre- and postoperatively, especially for parameters of gait. However, only gait time was both pre- and postoperatively significantly correlated to the KSS function score. The moderate correlations (Spearman's rho range 0.45–0.74) between the PROMs and performance-based measures may highlight the inherent link yet relative independence of the two measurement tools. PROMs are generic measures of outcome and include items across a broad range of health aspects. Therefore, PROMs may lack the level of detail in terms of relevance to any specific dimension measured with performance-based measures which perhaps accounts for the lower responsiveness and weaker correlations. The results of this study suggest that patients seem to score their physical performance best by reporting about their walking abilities, which may be best captured by the KSS function score. The study has some limitations that merit attention when interpreting the results. The cohort of patients undergoing a primary TKA in this study showed good outcomes of PROMs in comparison to previous findings [5,32], and there is a potential that the responsiveness of the outcome measures is slightly overstated compared to other populations. This might be due to the rather good physical status of this study's cohort as 8/20 patients were ASA class 1 and 10/20 patients had no other joint affected by OA (Charnley class A). Preoperative ASA
physical status score and Charnley classification have demonstrated to influence postoperative functional outcome following TKA [16,17]. Postoperative physical therapy was not standardized which could account for individual differences. Furthermore, as the population does not represent a consecutive case series, selection bias could be a confounder. Other limitations include the small sample size and the lack of more follow-up moments. Nonetheless, the prospective nature of the study, the use of a standardized performance-based test with novel IMU based outcome measures and the comparison with conventional patient-reported outcome measures are a strength and provide more insight into postoperative TKA recovery for comprehensive assessment of functional outcome. 5. Conclusion PROMs and physical performance-based outcome measures are both responsive to changes one year after TKA, but the moderate correlations may suggest that they measure a different dimension of outcome. Performance-based measures are necessary to objectively assess the patients' actual physical function that is not captured by PROMs and to fully characterize the changes in physical function of patients after TKA, particularly in younger or high demand patients where ceiling effects may be a problem. A physically challenging task, such as stepping onto a block, seems to capture the functional disabilities of patients with knee OA and the functional improvement after TKA better than a less challenging task, such as walking a short distance. To capture the variety of a patient's functional disability, we suggest that a routine clinical performance-based test should equally consist of a variety of daily life activities. Moreover, we suggest that a performance-based test should not only represent time-related measures but comprise a biomechanical approach capturing the underlying mechanisms causing functional impairment, to guide more targeted rehabilitation and to improve overall functional outcome. Conflict of interest statement None. References [1] Stevens-Lapsley JE, Schenkman ML, Dayton MR. Comparison of self-reported knee injury and osteoarthritis outcome score to performance measures in patients after total knee arthroplasty. PMR 2011;3(6):541–9 [quiz 549]. [2] Nilsdotter AK, Toksvig-Larsen S, Roos EM. Knee arthroplasty: are patients' expectations fulfilled? A prospective study of pain and function in 102 patients with 5year follow-up. Acta Orthop 2009;80(1):55–61. [3] Bolink SA, et al. Inertial sensor motion analysis of gait, sit–stand transfers and stepup transfers: differentiating knee patients from healthy controls. Physiol Meas 2012; 33(11):1947–58. [4] Vissers MM, et al. Psychological factors affecting the outcome of total hip and knee arthroplasty: a systematic review. Semin Arthritis Rheum 2012;41(4):576–88. [5] Senden R, et al. The importance to including objective functional outcomes in the clinical follow up of total knee arthroplasty patients. Knee 2011;18(5):306–11.
S.A.A.N. Bolink et al. / The Knee 22 (2015) 618–623 [6] Ornetti P, et al. Gait analysis as a quantifiable outcome measure in hip or knee osteoarthritis: a systematic review. Joint Bone Spine 2010;77(5):421–5. [7] Terwee CB, et al. Self-reported physical functioning was more influenced by pain than performance-based physical functioning in knee-osteoarthritis patients. J Clin Epidemiol 2006;59(7):724–31. [8] Boonstra MC, De Waal Malefijt MC, Verdonschot N. How to quantify knee function after total knee arthroplasty? Knee 2008;15(5):390–5. [9] Dobson F, et al. Measurement properties of performance-based measures to assess physical function in hip and knee osteoarthritis: a systematic review. Osteoarthritis Cartilage 2012;20(12):1548–62. [10] van den Dikkenberg N, et al. Measuring functional abilities of patients with knee problems: rationale and construction of the DynaPort knee test. Knee Surg Sports Traumatol Arthrosc 2002;10(4):204–12. [11] Dobson F, et al. OARSI recommended performance-based tests to assess physical function in people diagnosed with hip or knee osteoarthritis. Osteoarthritis Cartilage 2013;21(8):1042–52. [12] Hjorth MH, et al. Block-step asymmetry 5 years after large-head metal-on-metal total hip arthroplasty is related to lower muscle mass and leg power on the implant side. Clin Biomech (Bristol, Avon) 2014;29(6):684–90. [13] Wylde V, et al. Assessing function in patients undergoing joint replacement: a study protocol for a cohort study. BMC Musculoskelet Disord 2012;13:220. [14] Wylde V, et al. Does measuring the range of motion of the hip and knee add to the assessment of disability in people undergoing joint replacement? Orthop Traumatol Surg Res 2014;100(2):183–6. [15] Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957;16(4):494–502. [16] Hooper GJ, et al. The relationship between the American Society Of Anesthesiologists physical rating and outcome following total hip and knee arthroplasty: an analysis of the New Zealand Joint Registry. J Bone Joint Surg Am 2012;94(12):1065–70. [17] Bjorgul K, Novicoff WM, Saleh KJ. Evaluating comorbidities in total hip and knee arthroplasty: available instruments. J Orthop Traumatol 2010;11(4):203–9. [18] Bellamy N, et al. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol 1988;15(12):1833–40.
623
[19] Insall JN, et al. Rationale of the Knee Society clinical rating system. Clin Orthop Relat Res 1989;248:13–4. [20] Unnanuntana A, et al. Performance-based tests and self-reported questionnaires provide distinct information for the preoperative evaluation of total hip arthroplasty patients. J Arthroplasty 2012;27(5):770–5 [e1]. [21] Senden R, et al. Acceleration-based gait test for healthy subjects: reliability and reference data. Gait Posture 2009;30(2):192–6. [22] Janssen WG, Bussmann HB, Stam HJ. Determinants of the sit-to-stand movement: a review. Phys Ther 2002;82(9):866–79. [23] Lebel K, et al. Inertial measures of motion for clinical biomechanics: comparative assessment of accuracy under controlled conditions — effect of velocity. PLoS One 2013;8(11):e79945. [24] Takeda R, et al. Gait posture estimation using wearable acceleration and gyro sensors. J Biomech 2009;42(15):2486–94. [25] Seel T, Raisch J, Schauer T. IMU-based joint angle measurement for gait analysis. Sensors (Basel) 2014;14(4):6891–909. [26] Wong WY, Wong MS. Trunk posture monitoring with inertial sensors. Eur Spine J 2008;17(5):743–53. [27] Gonzalez RC, et al. Real-time gait event detection for normal subjects from lower trunk accelerations. Gait Posture 2010;31(3):322–5. [28] Bautmans I, et al. Reliability and clinical correlates of 3D-accelerometry based gait analysis outcomes according to age and fall-risk. Gait Posture 2011;33(3):366–72. [29] Boonstra MC, et al. Sit-to-stand movement as a performance-based measure for patients with total knee arthroplasty. Phys Ther 2010;90(2):149–56. [30] Hossain FS, et al. A performance based patient outcome score for active patients following total knee arthroplasty. Osteoarthritis Cartilage 2013;21(1):51–9. [31] Mizner RL, et al. Measuring functional improvement after total knee arthroplasty requires both performance-based and patient-report assessments: a longitudinal analysis of outcomes. J Arthroplasty 2011;26(5):728–37. [32] Khanna G, et al. Comparison of patient-reported and clinician-assessed outcomes following total knee arthroplasty. J Bone Joint Surg Am 2011;93(20):e117 [(1)–(7)].
