ORTHOPAEDIC SURGERY ANZJSurg.com
Subjective and objective outcome measures after total knee replacement: is there a correlation? Christy Graff,* Erik Hohmann,† Adam L. Bryant*‡ and Kevin Tetsworth‡§ *Musculoskeletal Research Unit, Central Queensland University, Rockhampton, Queensland, Australia †Centre for Health, Exercise and Sports Medicine, Department of Physiotherapy, Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne, Victoria, Australia ‡School of Medicine, The University of Queensland, Brisbane, Queensland, Australia and §Department of Orthopaedic Surgery, Royal Brisbane and Women’s Hospital, Brisbane, Queensland, Australia
Key words laxity, outcomes, strength, total knee replacement. Correspondence Dr Christy Graff, Musculoskeletal Research Unit, Central Queensland University, Canning Street, Rockhampton, QLD 4700, Australia. Email:
[email protected] C. Graff MBBS, MHMSc, FRCS; E. Hohmann MBChB, FRCS, FRCS (Tr&Orth), MD, PhD; A. L. Bryant BSc (Hons), PhD; K. Tetsworth MD, FRACS. Accepted for publication 3 July 2016. doi: 10.1111/ans.13708
Abstract Background: Although various methods for quantifying outcomes following total knee replacement (TKR) are used, there are few studies of the relationships between patient reported scores and functional tests. This paper aims to assess the relationships between commonly used outcome scores after TKR through a prospective cohort study. Methods: Twenty-four patients who had undergone unilateral TKR were assessed using four patient-reported outcome scores as well as objective measurements of knee laxity, quadriceps muscle strength and the Timed Up and Go Test. All scores and measures were correlated using the Pearson product moment correlation coefficient using the lower onesided 95% confidence interval. A level of significance of P < 0.05 was selected. Results: The Timed Up and Go Test was the only objective measure to demonstrate a statistically significant correlation (r = −0.557 to −0.770, P = 0.0001–0.005) with patientreported scores. Conclusion: A comprehensive assessment of outcomes after TKR requires both subjective and objective assessments. Walking ability and speed are important to the TKR patient and are representative of their pain and function.
Introduction The initial total knee replacement (TKR) procedures were initially judged only by pain relief – the first generation implants were without precedent.1 However, as arthroplasty has matured, we have collectively moved past the era of technical innovation and now have a greater interest in assessment and accountability.2,3 Determining data quality after any surgical intervention has become a very important consideration, particularly for elective procedures such as joint replacement. Only by determining the extent of the benefit using validated outcome measurements can we make informed clinical decisions regarding the costs and risks to our patients. Clinical results are currently assessed using both objective measures, such as range of movement or radiographic alignment, and subjective measures via patient-reported validated outcome instruments. While there are many methods for assessing outcomes after TKR, none are considered to be a universally accepted standard.3–6 In comparative studies of patient-reported outcome scores for TKR, the most widely used was the American Knee Society Score (KSS). © 2016 Royal Australasian College of Surgeons
The most frequently recommended were the Oxford Knee Score (OKS), the Western Ontario and McMasters Universities Osteoarthritis Index (WOMAC) and the Knee Injury and Osteoarthritis Outcomes Score (KOOS).4,5,7 In addition, the OKS has been used in National Joint Registries in several countries, including England, New Zealand and Sweden.8 Validated, general quality of life metrics include the Short Form 36 and the abbreviated version (Short Form 12 (SF12)).5,7 Considering these alternatives, patient-reported outcome measures in our study included the KSS, OKS, KOOS and SF12. In addition, several authors advocate using objective measures such as knee range of motion, knee stability, muscle strength and functional tests (such as walking speed and stair climbing) to assess outcome because improving knee function is a prime concern for TKR patients.9–11 Recently, researchers have demonstrated that dependence on patient-based outcome scores alone in the perioperative period after TKR results in an overestimation of their function.10–12 It has been suggested that patient-reported measures of activity have poor correlation to performance-based ANZ J Surg 86 (2016) 921–925
Graff et al.
922
measures.12,13 Unfortunately, insufficient data exist on the relationship between these two types of outcome measures.13–15 The perception of an individual’s functional levels and satisfaction coupled with objective measures of function may provide valuable insight into rehabilitation strategies and outcomes and ultimately provide a foundation for evidence-based practice.15,16 Given these considerations, the aim of this study was to investigate the potential relationships between four validated and widely used patient-reported questionnaires (KSS, OKS, KOOS and SF12) and three objective outcome measures (muscle strength, knee laxity and the Timed Up and Go Test (TUG)). We hypothesized that there would be a positive correlation between patient-reported outcome scores and both muscle torque and TUG, but not knee laxity.
Methods
12 categories. The worst score was 60 (the worst outcome accrues five points). The KOOS contains 42 items in five separately scored subscales: pain, other knee symptoms, activity of daily living (ADL), function in Sport and Recreation and knee-related quality of life. Each item has a maximum of 4 points and a minimum score of 0. These were then summed, and a score out of 100 was generated for each subscale, with 100 being the best outcome and 0 being the worst as per the original authors.17 The subscales were not designed to be used cumulatively. The KSS was designed as two different scores. The KSS is made up of pain, range of motion and stability with a maximum of 100 points. The function score is comprised of walking distance, stair climbing and walking aids with a maximum of 100 points. The SF12 is divided into a physical component score and a mental component score, with a maximum score of 100 points each.
Study participants Twenty-four patients were recruited from the Orthopaedics Outpatient Department. The inclusion criteria included the ability to give consent, primary unilateral TKR at least 12 months prior, no contralateral TKR and no surgery or trauma to the contralateral limb within the prior 12 months. Exclusion criteria included subject refusal or surgeon assessment that the subject is unable to participate as well as medical or surgical conditions that prevented the subject from mobilizing. Ethical clearance was obtained from both the University Human Ethics Research Review Panel and the Health District.
Surgical technique All patients had a cemented Duracon Cruciate Retaining Total Knee Replacement (Stryker Corporation, Kalamazoo, MI, USA) for osteoarthritis. All TKRs were performed through a medial parapatellar approach; 10 of these were navigated. None of the patellas were resurfaced. All patients received routine follow-up with standard radiographs at 6 weeks, 3, 6 and 12 months, and annually thereafter.
Objective outcome measures Knee laxity was tested using the KT1000 arthrometer (MEDmetric, San Diego, CA, USA) according to the User’s Guide for the Knee Ligament Arthrometer.18 As per the protocol, both lower limbs were positioned in the same degree of flexion and rotation with the knee flexed to 20 5 (Fig. 1). The non-operated knee was tested first followed by the operated knee. With the arthrometer positioned on the anterior aspect of the leg, an upward force was applied through the handle. The audiotone signalled when 15- and 20-lb forces were applied, and the displacement was read in mm from the dial indicator. This was repeated with a 20-lb force directed posteriorly. The arthrometer was zeroed prior to each reading, and the readings were only accepted if the indicator needle returned to zero when the handle tension was released. This was repeated three times, and the scores were averaged. Muscle strength was assessed using a Biodex Isokinetic Dynamometer (System 3; Biodex, Shirley, NY, USA). The non-involved leg was tested first to calculate symmetry indices. The isokinetic eccentric strength was tested at 120 /s. Each session began with a practice trial serving both as a warm up and a familiarization
Experimental protocol Participants attended the Human Movement University Laboratory (Central Queensland University, Rockhampton, Queensland, Australia) for testing in the gait laboratory in a defined order. All subjects were offered an introductory familiarization session 1 day prior to their testing; however, all subjects chose to familiarize themselves with the equipment on the day of testing. One investigator (CG), who was skilled in all instruments and was trained in orthopaedics, tested all subjects.
Subjective outcome measures Subjective responses of TKR subjects were derived using the OKS, KOOS, SF12 and KSS. The OKS and SF12 were sent as postal questionnaires; the KOOS and KSS were completed during the testing session. The OKS has 12 items, for which the best total score was 12 – the best outcome accrues one point in each of the
Fig. 1. Testing knee laxity using the KT1000 arthrometer according to the User’s Guide for the Knee Ligament Arthrometer. © 2016 Royal Australasian College of Surgeons
923
Outcome measures after TKR
exercise. A 30-s rest interval between all sets was used to avoid fatigue. Each subject then performed one set of three maximal extension and flexion repetitions. Verbal encouragement helped maximize performance. Subjects were rested for 5 min after the Biodex testing prior to performing the TUG.19 Subjects were asked to first rise comfortably from a chair and then walk as fast as they could at a comfortable pace (without running) for a distance of 3 m and then return to a sitting position in the chair. Subjects were allowed to use the walking devices that they used normally. Subjects completed this task three times with a rest period of 30 s between – we used the fastest time.
Statistical analyses Power calculation A power calculation for sample size was performed and was based on Wilk et al.20 An alpha level of 0.05 and power of 80% suggested that 23 complete datasets were needed to achieve adequate statistical power. Descriptive statistics, including the mean and standard deviation, were calculated. Correlational analyses were then performed to examine the relationships among the knee rating scales and between the knee rating scales and objective measures. All scores and measures were correlated using the Pearson product moment correlation coefficient. The coefficient of correlation ‘r’ was interpreted according to Cohen (1988):21 0.0–0.3 (weak), 0.3–0.5 (moderate) and >0.5 (strong). The lower one-sided 95% confidence interval was calculated for each correlation coefficient. Significance values of P < 0.05 were selected in all analyses to limit the chance of type I error to 5%. All analyses were conducted using SPSS (Version 12.0.1; SPSS Inc., Chicago, IL, USA) for Windows.
Results The mean age was 68.0 6.88 years (range: 54–80 years). There were 13 males with a mean age of 65.5 5.08 years (range: 54–80 years) and 11 females with a mean age of 68.72 8.26 years (range: 62–77 years). The mean body mass index of the subjects (weight (kg)/(height (m))2) was 31.5 5.88 (range: 19.8–42.1). The mean time from surgery was 27.5 11.7 months (12–58 months). Nine patients had osteoarthritis of the contralateral knee. Two subjects did not complete the SF12, and two subjects did not complete the OKS. We could not contact subjects to complete the questionnaires. One subject was unable to complete the Biodex muscle testing because of unrelated illness. Table 1 summarizes the results for all outcome scores and Table 2 summarizes the results for all objective measures. Table 3 summarizes the correlations between subjective outcome measures and the TUG. Strong and highly significant correlations (r = −0.557 to −0.770, P = 0.0001–0.005) were observed between TUG and the knee rating scores (OKS, KSS knee and four items of the KOOS). There were no other significant correlations between the objective measures and the knee rating scores. © 2016 Royal Australasian College of Surgeons
Table 1 Knee rating systems scores Knee rating score
Score
Standard deviation
Confidence interval
OKS KOOS symptoms KOOS pain KOOS ADL KOOS sports KOOS QOL KSS knee KSS function SF12 PCS SF12 MCS
23.1 80.2 79.7 78.8 39.6 63.2 75.1 65.38 37.89 51.24
10.9 17.3 19.8 19.9 33.8 30.9 19.41 17.69 10.46 10.72
4.6 6.9 7.9 7.94 13.5 12.4 7.69 7.08 4.37 4.47
ADL, activity of daily living; KOOS, Knee Injury and Osteoarthritis Outcomes Score; KSS, Knee Society Score; MCS, mental component score; OKS, Oxford Knee Score; PCS, physical component score; QOL, quality of life; SF12, Short Form 12.
Table 2 Objective measures Objective measure
Score
Standard deviation
Confidence interval
Muscle torque (O) (Nm/kg) Muscle torque (NO) (Nm/kg) Muscle torque symmetry index Knee laxity (O) (mm) Knee laxity (NO) (mm) Knee laxity difference (mm) TUG (s)
113.01 113.69 1.00 3.83 4.13 1.96 7.89
9.04 8.37 0.15 2.03 2.15 1.27 2.66
3.69 3.42 0.06 1.5 1.65 0.51 1.06
NO, non-operated limb; O, operated limb; TUG, Timed Up and Go Test.
Table 3 Relationship between TUG and outcome scores Knee scores
OKS-TUG KSS knee-TUG KSS function-TUG KOOS symptoms-TUG KOOS pain-TUG KOOS ADL-TUG KOOS sports-TUG KOOS QOL-TUG PCS-TUG MCS-TUG
Pearson correlation
P-levels
Critical R value
0.651 −0.770 −0.513 −0.680 −0.679 −0.723 −0.557 −0.686 −0.392 −0.424
0.001 0.000 0.010 0.000 0.000 0.000 0.004 0.000 0.058 0.039
0.404 (r22) 0.404 (r22) 0.404 (r22) 0.404 (r22) 0.404 (r22) 0.404 (r22) 0.404 (r22) 0.404 (r22) 0.413 (r21) 0.413 (r21)
ADL, activity of daily living; KOOS, Knee Injury and Osteoarthritis Outcomes Score; KSS, Knee Society Score; MCS, mental component score; OKS, Oxford Knee Score; PCS, physical component score; QOL, quality of life; TUG, Timed Up and Go Test.
Discussion The most important finding in this study was the strong correlation between the TUG and most of the subjective rating scores. This would suggest that walking ability and speed are important to the TKR patient and are representative of their pain and function. Walking velocity has already been shown to have a relationship with knee rating scores and patient satisfaction after TKR.13,22 Rossi et al.13 reported statistically significant (P < 0.05) correlations between the TUG and the aggregate WOMAC score (r = 0.59, r2 = 0.553) and the physical function score (r = 0.63,
Graff et al.
924
r2 = 0.553) at 10–26 months post-operatively. Tekin et al.23 reported that improvements in pain and gait were the two most commonly expected outcomes after TKR. It is not surprising that the TUG was correlated with the patient-based outcome scores in the current research because walking speed is a predictor of disability, dependence and social networks in older populations.24,25 No significant correlations were identified between the peak muscle torque and the knee rating scores. In contrast, Silva et al.14 demonstrated a positive correlation (r = 0.57, P = 0.04 and r2 = 0.349) between isometric peak quadriceps torque and the KSS function score in a population of 19 TKR patients with ‘clinically well functioning’ TKRs. Mizner et al.12 also demonstrated a correlation between isometric quadriceps strength of the involved limb and a knee rating score (KOOS-ADL) both pre-operatively (r = 0.28, P < 0.05 and r2=0.195) and 1 year post-operatively (r = 0.26, P < 0.05 and r2 = 0.195), but not 1 month post-operatively. It is difficult to compare our results to these previous findings due to differences in methodology. The reported correlations in both papers were weak (i.e. r = 0.57, P = 0.05 and r2 = 0.349;14 r = 0.26, P < 0.05 and r2 = 0.19512). Both authors used isometric peak torque. In this study, isokinetic muscle strength was used rather than isometric peak torque – it did not correlate with patientreported outcomes. There were no observed correlations between knee laxity and the knee rating scores. Dejour et al.26 could not identify a relationship between knee laxity after TKR and the WOMAC, KSS or SF12. Similarly, Warren et al.27 could not find an association between anteroposterior laxity and patients’ satisfaction or pain relief using the KSS. More recently, Seon et al.28 used the Telos arthrometer (Austin & Associates Inc., Fallston, MD, USA) and stress radiographs to correlate mean anteroposterior laxity with the Hospital for Special Surgery score and WOMAC scores. They reported no differences in either group with regard to the Hospital for Special Surgery score and WOMAC total score but found that the stable group (
