You are on page 1of 34

Accepted Manuscript

Validity and Responsiveness of the 2-Minute Walk Test for Measuring Functional
Recovery after Total Knee Arthroplasty

Aasis Unnanuntana, MD, Pakpoom Ruangsomboon, MD, Worawut Keesukpunt, MD

PII: S0883-5403(18)30033-0
DOI: 10.1016/j.arth.2018.01.015
Reference: YARTH 56336

To appear in: The Journal of Arthroplasty

Received Date: 21 November 2017


Revised Date: 28 December 2017
Accepted Date: 8 January 2018

Please cite this article as: Unnanuntana A, Ruangsomboon P, Keesukpunt W, Validity and
Responsiveness of the 2-Minute Walk Test for Measuring Functional Recovery after Total Knee
Arthroplasty, The Journal of Arthroplasty (2018), doi: 10.1016/j.arth.2018.01.015.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT

Validity and Responsiveness of the 2-Minute Walk Test for Measuring Functional

Recovery after Total Knee Arthroplasty

Aasis Unnanuntana, MD1; Pakpoom Ruangsomboon, MD1; Worawut Keesukpunt, MD2

PT
1
Department of Orthopaedic Surgery, Faculty of Medicine Siriraj Hospital, Mahidol

RI
University, Bangkok, Thailand

SC
2
Department of Orthopaedics, Rayong Hospital, Rayong, Thailand

U
Conflict of interest statement: We have no conflicts.
AN
Corresponding author:

Aasis Unnanuntana, MD
M

Associate Professor of Orthopedic Surgery


D

Department of Orthopaedic Surgery


TE

Faculty of Medicine Siriraj Hospital, Mahidol University

2 Wang Lang Road, Bangkok Noi


EP

Bangkok 10700, Thailand

Tel: (+66) 2-419-7968; Fax: (+66) 2-412-8172


C

E-mail: uaasis@gmail.com
AC
1
ACCEPTED MANUSCRIPT
1 Validity and Responsiveness of the 2-Minute Walk Test for Measuring Functional

2 Recovery after Total Knee Arthroplasty

3 ABSTRACT

4 Background: The 2-minute walk test (2mwt) is a performance-based test that evaluates

5 functional recovery after total knee arthroplasty (TKA). This study evaluated its validity

PT
6 compared with the modified Western Ontario and McMaster Universities Osteoarthritis Index

RI
7 (WOMAC), Oxford Knee Score (OKS), modified Knee Score (modified KS), Numerical

8 Pain Rating Scale (NPRS) and Timed Up and Go test (TUG), and its responsiveness in

SC
9 assessing functional recovery in TKA patients.

10 Methods: This prospective cohort study included 162 patients undergoing primary TKA

11
U
between 2013 and 2015. We used patient-reported outcome measures (modified WOMAC,
AN
12 OKS, modified KS, NPRS) and performance-based tests (2mwt and TUG) at baseline and 3,
M

13 6 and 12 months postoperatively. The construct validity of 2mwt was determined between the

14 2mwt-distances walked and other outcome measurements. To assess responsiveness, effect


D

15 size and standardized response mean were analyzed. Minimal clinically important difference
TE

16 (MCID) of 2mwt at 12 months after TKA was also calculated.

17 Results: All outcome measurements improved significantly from baseline to 3, 6 and 12


EP

18 months postoperatively. Bivariate analysis revealed mild to moderate associations between


C

19 the 2mwt and modified WOMAC function subscales, and moderate to strong associations
AC

20 with OKS. Mild to moderate correlations were found for pain and stiffness between 2mwt

21 and other outcome measurements. The effect size and standardized response mean at 12

22 months were large, with an MCID of 12.7 meters.

23 Conclusion: 2mwt is a valid performance-based test with responsiveness properties. Being

24 simple and easy to perform, it can be used routinely in clinical practice to evaluate functional

25 recovery after TKA.


2
ACCEPTED MANUSCRIPT
26 Keywords: 2-minute walk test, Timed Up and Go test, validity, responsiveness, total knee

27 arthroplasty, functional recovery, osteoarthritis knee

PT
RI
U SC
AN
M
D
TE
C EP
AC
3
ACCEPTED MANUSCRIPT
28 Introduction

29 As total knee arthroplasty (TKA) continues to evolve over time, an increasing number

30 of outcome measures targeting the functional status of patients with knee osteoarthritis (OA)

31 and those undergoing a TKA have been developed [1-3]. These measures include patient-

32 reported outcome measures (PROMs) and performance-based tests, with PROMs being

PT
33 endorsed and applied more often [4]. However, some studies have encouraged the use of

RI
34 performance-based tests as important tools for capturing the actual improvement in functional

35 status as PROMs often overestimate the ability of patients to ambulate after total hip

SC
36 arthroplasty [5,6]. For example, Parent et al. [7] demonstrated that although patients with

37 total knee arthroplasty showed dramatic improvements at 2 months, as assessed by the

38
U
Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) function
AN
39 subscale, there were no significant changes when measured with performance-based tests
M

40 assessing gait speed and stair ascent duration. Therefore, clinicians should be cautious when

41 interpreting findings from studies providing PROM scores only. Given that PROMs may
D

42 overestimate the functional status after joint arthroplasty, some authors have advocated the
TE

43 use of performance-based tests as another outcome measure, recommending their use during

44 the preoperative period for baseline measurements, followed by subsequent comparison with
EP

45 postoperative measurements [2,8].


C

46 The 6-minute walk test and the Timed Up and Go test (TUG) are among the most
AC

47 commonly reported performance-based tests used to assess the function of the lower

48 extremities [2,3,9]. Although the 6-minute walk test has been shown to be a valid and reliable

49 outcome measure [3], it may be difficult to perform routinely, especially during the

50 preoperative period with patients using walking aids. An alternative walking test is the 2-

51 minute walk test. This is a shorter, more clinically practical, walking test that is commonly

52 used in patients with cardiopulmonary diseases [10,11]. Both the 2-minute walk test and
4
ACCEPTED MANUSCRIPT
53 TUG are simple and quick, require minimal staff and equipment, and can be applied in a

54 short period of time as part of a routine follow-up after a TKA.

55 To utilize the 2-minute walk test in TKA patients routinely, it has to be reliable, valid,

56 and responsive in this particular population, since a measure that has been shown to be valid

57 in one clinical context may not be valid in a different context. Although the validity of the 2-

PT
58 minute walk test has previously been reported in patients undergoing total hip arthroplasty

RI
59 [6,12], no studies have validated and documented the longer-term outcomes and

60 responsiveness of this particular walking test in TKA patients. Responsiveness is defined by

SC
61 the ability to detect changes in a clinical condition. Measures of responsiveness have

62 commonly been reported as effect size, standardized response mean (SRM), and minimal

63
U
clinically important difference (MCID). This provides a threshold that permits a judgement to
AN
64 be made of whether a difference between two outcome scores in a population of patients
M

65 represents a true change in the outcome being assessed, or can be attributed to measurement

66 inaccuracies [13-15].
D

67 The objective of this study was to characterize the utility of the 2-minute walk test in
TE

68 a TKA population. The specific study goals were: 1) describe the changes of a 2-minute walk

69 test at 3, 6, and 12 months postoperatively; 2) evaluate the validity of the 2-minute walk test
EP

70 by comparing this test with other validated measures that are commonly used to assess
C

71 mobility in TKA patients; and 3) measure the responsiveness of the 2-minute walk test at 12
AC

72 months postoperatively by calculating the effect size, SRM and the minimal detectable

73 change of score.

74

75 Materials and Methods

76 The protocol and consent forms used in this study were approved by the XXX

77 Institutional Review Board (SIRB), Faculty of Medicine, XXX Hospital, XXX University.
5
ACCEPTED MANUSCRIPT
78 This study was registered in the ClinicalTrials.gov database via the Protocol Registration and

79 Results System (NCT02156453). Patients treated by a single surgeon and scheduled for

80 primary knee arthroplasty between April 2013 and October 2015 were prospectively

81 included. Patients were eligible if they were between 40 and 100 years of age and were

82 medically stable. They were excluded if they had received unicompartmental knee

PT
83 arthroplasty; had postoperative complications such as an infection or fracture; had undergone

RI
84 a complex surgical procedure which required bone grafting, a metal augment, a sleeve or a

85 stem during the operation; or had been diagnosed with a neurological disorder that caused

SC
86 gait disturbances. Patients who did not return PROMs or were unable to understand verbal

87 and written instructions were also excluded. For those who had been diagnosed with bilateral

88
U
OA knees and were scheduled for a staged bilateral procedure within a year, we collected
AN
89 data from the second TKA. The recruitments and enrollments were conducted when the
M

90 patients visited the hospital during their hospital admission. At enrollment, patients

91 completed all PROMs and were asked to execute the performance-based tests.
D

92 Operative procedure and postoperative care


TE

93 All surgeries were performed by a single surgeon (AU) using a medial parapatellar

94 approach. The prostheses used in this study were either the Press Fit Condylar Sigma (PFC
EP

95 Sigma) posterior stabilized design (DePuy Synthes, Warsaw, IN) or the Attune posterior
C

96 stabilized design (DePuy Synthes, Warsaw, IN). The patella was selectively resurfaced, based
AC

97 on the intraoperative findings. If the cartilage on the patella was generally preserved with

98 adequate patellofemoral congruency and there was no history of crystalline and inflammatory

99 synovitis, the patellar was not resurfaced [16]. All components were cemented. Following the

100 TKA, all participants underwent a standardized 4- to 5-day in-hospital care protocol.

101 Generally, all patients were mobilized out of bed and started walking, as tolerated, on the first

102 postoperative day. The postoperative pain was controlled with a combination of oral
6
ACCEPTED MANUSCRIPT
103 analgesics, including a nonsteroidal analgesic drug, tramadol, tylenol and oral opioids, as

104 needed. Intravenous morphine was given when the patient’s pain score was > 5. The goal was

105 to keep pain at 2 to 3 out of 10, as measured by the Numerical Pain Rating Scale (NPRS).

106 The intravenous fluid and foley catheter were typically removed on the second postoperative

107 day. All patients were discharged to their home. Patients were instructed to do simple knee

PT
108 exercises and to walk using a walker a few times a day for 2–3 weeks after discharge. One

RI
109 month after surgery, they were instructed to continue doing a range of motion exercises and

110 quadriceps exercises, and practice walking with or without an assistive walking device for

SC
111 approximately 20–30 minutes each time, twice a day.

112 Testing procedures

113
U
All patients underwent testing at 4 different periods in this prospective cohort
AN
114 investigation: 1 day before surgery (to establish the baseline data), and 3, 6 and 12 months
M

115 postoperatively. Patients were asked to complete all PROMs before performing the 2-minute

116 walk test and TUG.


D

117 Patient-reported outcome measures


TE

118 Modified WOMAC

119 WOMAC is a disease-specific questionnaire, which is valid, reliable, and sensitive to


EP

120 change in the health status of patients with hip and knee OA [1,17]. The original WOMAC is
C

121 composed of 24 items grouped into 3 dimensions: pain (5 items), stiffness (2 items), and
AC

122 physical function (17 items). However, in this study, we used the 22-item, modified Thai

123 WOMAC index. Because of some cross-cultural differences, a Thai index had been modified

124 from the original index (the 17-item functional subscale was decreased to 15 items) and had

125 been validated by previous investigators for OA knees [18]. The test-retest reliability of this

126 modified WOMAC had correlation coefficients ranging from 0.65–0.71, with its internal
7
ACCEPTED MANUSCRIPT
127 consistency ranging from 0.85–0.97 [18]. Similar to the original WOMAC, higher scores

128 indicate a worsened function of the knee, while lower scores indicate an improved function.

129 Oxford Knee Score (OKS)

130 OKS is a 12-item, patient-reported questionnaire specifically designed and developed

131 to assess pain and function after a TKA [19]. Each question is based on one modality, which

PT
132 can be classified into 5 grades of severity, ranging from 0 (the worst) to 4 (the best). The final

RI
133 score therefore ranges from the worst (0) to the best (48) patient outcome. This tool has been

134 proved to be valid, clinically meaningful, and easy to perform [20]. In addition, OKS has

SC
135 been shown to have a good correlation with patients’ postoperative satisfaction levels [21].

136 Modified Knee Score

137
U
This scoring system was modified from the original Knee Society Clinical Rating
AN
138 system by Insall in 1993 [22]. The system is based on 3 categories: pain, range of motion and
M

139 stability. The total of the individual scores given for each category is then reduced by the

140 total of the scores given for extension lag, flexion contracture, malalignment and pain at rest,
D

141 resulting in an overall modified knee score. The modified knee score is based on a total of
TE

142 100 points, with higher scores indicating an improved function of the knee, and lower scores

143 indicating a worsened function.


EP

144 Numeric Pain Rating Scale (NPRS)


C

145 The knee pain experienced by the patients was recorded using the NPRS. Subjects
AC

146 quantified their pain intensity on a scale ranging from 0 (no pain) to 10 (extreme pain, or pain

147 as bad as it can be). NPRS is efficient for use in clinical practice and has been demonstrated

148 to have a good test-retest reliability in different population [23,24].

149 Performance-Based Tests

150 Functional evaluation was conducted using 2 performance-based tests: the 2-minute

151 walk test and TUG test. The results of these tests were recorded preoperatively (to establish
8
ACCEPTED MANUSCRIPT
152 the baseline data) and postoperatively (at the 3-, 6- and 12-month follow-ups) by a research

153 assistant.

154 Two-Minute Walk Test

155 Patients were instructed to walk for 2 minutes at their normal pace up and down a

156 designated corridor, turning around at each end of the corridor without stopping [25]. They

PT
157 were permitted to use walking aids if they wished. The results were recorded as total distance

RI
158 walked in meters.

159 Timed Up and Go Test

SC
160 Patients were instructed to rise from a high-seated chair, walk at a safe and

161 comfortable pace to a mark 3 meters away, and return to a sitting position with their backs

162
U
against the chair [2]. Patients were permitted to use their arms when rising from and returning
AN
163 to a seated position. A stopwatch was used to measure the time to complete this activity to
M

164 the nearest one-tenth of a second. Patients were asked to perform this task 3 times, and the

165 average time was used for analysis.


D

166 Demographic data and clinical information were collected at enrollment. Specifically,
TE

167 the patients’ physical status was measured with the American Society of Anesthesiologists’

168 (ASA) classification, a widely-used grading system to assess the preoperative health of
EP

169 surgical patients. The ASA score, a subjective rating of a patient’s overall health, is presented
C

170 as five classes (I to V) [26]. In class I, the patients are completely healthy and fit. Class II
AC

171 defines patients with mild systemic disease, whereas Class III is for patients with a severe

172 systemic disease that is not incapacitating. Class IV patients have an incapacitating disease

173 which is a constant threat to life, while Class V is reserved for moribund patients who are not

174 expected to live longer than 24 hours, with or without surgery [27]. Active back pain and the

175 number of painful joints of the lower extremities were also recorded; the results were

176 reported as the total number of additional painful joints in each patient. We also recorded the
9
ACCEPTED MANUSCRIPT
177 use of walking aids, categorized as none (walking without any assistive device) or requiring a

178 cane, walker or wheelchair.

179 Sample size calculation

180 A power analysis was conducted using nQuery Advisor 6.0. The results of a previous

181 investigation [28] found that the effect size of the 6-minute walk test was approximately 0.66.

PT
182 A power analysis conducted a priori determined that a minimum of 128 subjects were

RI
183 required to establish a minimum effect size of difference of 0.25, with a 2-sided alpha level

184 of 0.05 and 80% power. Since recruitment was increased by 20% to compensate for loss to

SC
185 follow-up, a total of 154 subjects were required for this study.

186 Data analysis

187
U
Data analyses were performed using SPSS Statistics for Windows, version 18.0
AN
188 (SPSS Inc., Chicago, Ill., USA). Data are presented as number and percentage, mean ±
M

189 standard deviation, or median [IQR] and range. Outcome scores for all instruments were

190 calculated at each follow-up visit. A p-value less than 0.05 was regarded as being statistically
D

191 significant for all tests.


TE

192 Validity

193 We determined the construct validity of the 2-minute walk test by calculating
EP

194 Pearson’s correlation coefficient between the scores of the 2-minute walk test and the scores
C

195 of the other outcome measurements (WOMAC, OKS, modified knee score, NPRS and TUG).
AC

196 Construct validity refers to the extent to which the scores on a particular measure relate to

197 other measures in a manner that is consistent with theoretically derived hypotheses

198 concerning the concepts being measured [29]. We evaluated the correlation between the data

199 of the 2-minute walk test and the scores from other outcome measures at all 4 time points

200 (baseline, and 3, 6 and 12 months after surgery). Correlation coefficients of 0.1 to 0.3 were

201 considered weak; 0.3 to 0.6, moderate; and > 0.6, strong [30].
10
ACCEPTED MANUSCRIPT
202 Responsiveness and minimal clinically important difference (MCID)

203 Responsiveness is the ability of a questionnaire to detect clinically important changes

204 over time. Changes in the scores of all instruments were calculated from baseline-data to data

205 recorded at the 12-month follow-up. To assess responsiveness, effect size and standardized

206 response mean were used. Effect size was defined as the mean change in the patient scores

PT
207 divided by the standard deviation (SD) of the baseline scores. Effect estimates were

RI
208 interpreted according to Cohen, as follows: a standardized response mean of 0.2 to 0.4 was

209 considered a small effect; 0.5 to 0.7, a moderate effect; and, ≥ 0.8, a large effect [31,32]. The

SC
210 standardized response mean was defined as the mean change in the patient scores divided by

211 the SD of the changed scores.

212
U
The minimal clinically important difference (MCID) of the 2-minute walk test was
AN
213 calculated using the distribution-based approach [33], which can be estimated using the
M

214 following formula: SD baseline [√(1-r)], where SD baseline is the standard deviation of the

215 baseline data of the 2-minute walk test, and r is the intraclass correlation coefficient of the 2-
D

216 minute walk test between the baseline and the 1-year follow-up.
TE

217

218 Results
EP

219 Of the total of 209 patients screened for this study, 162 were enrolled. The reasons for
C

220 exclusion were: patients who had received unicondylar knee arthroplasty (n=17); those who
AC

221 had undergone complex total knee arthroplasty (n=24); those with immediate postoperative

222 complications (n=5); and those who unable to follow verbal or written instructions (n=1).

223 With 5 patients also lost to follow-up during the 12-month study period, 157 subjects

224 remained in the study (Figure 1). The descriptive statistics on patient demographics and

225 clinical characteristics are at Table 1. The mean age of the included patients was 69.6 years,

226 with most being female (87.2%). Twenty-four patients (15.3%) had bilateral simultaneous
11
ACCEPTED MANUSCRIPT
227 TKAs. The majority of patients were ASA class II (79.0%). In the preoperative period, 78

228 patients (49.7%) walked without a gait aid. A majority of subjects did not have active back

229 pain or other joint pain of the lower extremities before the procedure (Table 1).

230 All outcome measurements had a symmetrical distribution. The baseline scores of

231 each preoperative outcome are at Table 2. The mean modified WOMAC pain, stiffness and

PT
232 function scores at baseline were 29.9, 11.2 and 93.3, respectively. The baseline OKS,

RI
233 modified knee score and NPRS were 16.6, 58.0 and 7.4, respectively. These scores improved

234 significantly from baseline to 3, 6 and 12 months, postoperatively (Figure 2). The mean

SC
235 distance walked in 2 minutes was 46.2 meters at baseline. The mean distance walked in 2

236 minutes improved significantly to 55.6, 62.5 and 66.1 meters at 3, 6 and 12 months after

237
U
surgery, respectively (p < 0.001; Figure 3A). The mean baseline TUG was 23.9 seconds,
AN
238 which subsequently decreased to 19.7, 17.1 and 15.6 seconds at 3, 6 and 12 months after
M

239 surgery, respectively. Like the 2-minute walk test results, the scores of the TUG improved

240 significantly at all time points (Figure 3B).


D

241 The correlation between the 2-minute walk test results and the scores of the other
TE

242 outcome measurements at each time point are at Table 3. A bivariate analysis revealed mild

243 to moderate associations between the 2-minute walk test and the modified WOMAC function
EP

244 subscale, and a moderate to strong association with the OKS. Conversely, there were no
C

245 associations between the 2-minute walk test and the modified knee score at 3 and 12 months,
AC

246 postoperatively (p-value > 0.05). The association between the 2-minute walk test and the

247 modified knee score at baseline and 6 months postoperatively was found to be weak, with an

248 r-value of 0.165 and 0.178, respectively. In the case of the correlation between the 2-minute

249 walk test and the outcome measurements for pain, a mild to moderate association was

250 observed between the 2-minute walk test and the modified WOMAC pain subscale at 3, 6

251 and 12 months postoperatively and the NPRS at all time points. In addition, there were mild
12
ACCEPTED MANUSCRIPT
252 associations between the 2-minute walk test and the modified WOMAC stiffness subscale at

253 6 and 12 months postoperatively. As for the correlation between the 2 performance-based

254 tests, the correlations between the 2-minute walk test and TUG were strong at all time points,

255 with r-values ranging from 0.781–0.831.

256 The responsiveness of each outcome measurement is at Table 4. The effect size and

PT
257 standardized response mean of the 2-minute walk test at 12 months after surgery were 1.01

RI
258 and 1.12, respectively. Based on the standardized response mean at 12 months after surgery,

259 all outcome measurements except the TUG had large effect estimates. The effect size and

SC
260 standardized response mean of the TUG at 12 months after surgery were 0.58 and 0.68,

261 respectively. Since the intraclass correlation coefficient of the 2-minute walk test between

262
U
baseline and the 1-year follow-up was 0.584 (based on the distribution-based approach), the
AN
263 MCID of the 2-minute walk test at 12 months after surgery was 12.7 meters. By using the
M

264 same equation, the MCID of the TUG at 12 months after surgery was approximately 9.5

265 seconds.
D

266
TE

267 Discussion

268 The level of impairment in patients with knee OA can be measured using validated
EP

269 PROM, which can be categorized into disease-specific questionnaires (such as WOMAC,
C

270 OKS and the modified knee score) and generic, quality-of-life questionnaires (such as the SF-
AC

271 36). Although PROM provides information on how individuals perceive their physical

272 function in their own environment, it does not describe their ability to perform a specific task

273 or action [34]. In addition, some investigators have demonstrated that the scores from

274 PROMs taken before and after arthroplasty can overestimate patients’ actual functional

275 capacity [5,7,35]. Therefore, evaluations that include both PROM and performance-based
13
ACCEPTED MANUSCRIPT
276 tests are important to provide a more comprehensive perspective of functional recovery after

277 a TKA.

278 To interpret data from any outcome measurement tool, it is a prerequisite that each

279 outcome measurement must be valid, reliable and responsive to changes in a patient’s

280 condition. To our knowledge, this is the first study to investigate the validity and

PT
281 responsiveness of the 2-minute walk test in patients who had undergone a TKA. We had

RI
282 hypothesized that the 2-minute walk test measures a different, but related, construct than that

283 measured by the currently available PROMs, and this hypothesis was supported by the

SC
284 findings that there were mild to moderate correlations between the 2-minute walk test and the

285 modified WOMAC, and a moderate to strong correlation between the 2-minute walk test and

286
U
OKS. This confirms that they do not measure exactly the same construct, but are merely
AN
287 related as they assess knee function in different ways. In contrast, there were strong
M

288 correlations between the 2-minute walk test and TUG (the r values ranged from 0.781–

289 0.831), yielding convergence validity between these 2 performance-based tests. Although
D

290 both the 2-minute walk test and TUG measure the function of the lower extremity, the 2-
TE

291 minute walk test assesses patients’ walking ability and cardiovascular fitness while walking

292 on a level surface, whereas TUG evaluates the ability of patients to adjust their body position
EP

293 with a voluntary movement, and assesses their strength for vertical and horizontal transitions
C

294 in their body’s position [25,36]. Thus, these 2 tests may be appropriate in different clinical
AC

295 settings, and a combination of these 2 performance-based tests is likely to provide a more

296 comprehensive assessment of function in TKA patients.

297 Unlike previous investigations which showed a moderate association between the

298 knee society score and some performance based tests [37], our study showed that there were

299 no associations between the 2-minute walk test and the modified knee score at 3 and 12

300 months postoperatively, whereas there were mild correlations between the 2-minute walk test
14
ACCEPTED MANUSCRIPT
301 and the modified knee score at baseline and 6 months post-TKA. This can be explained by

302 the fact that the modified knee score that was used in this study measured only pain, the

303 range of motion, and stability. There was no functional domain in this modified knee score.

304 Similar to the modified knee score, the correlations between the 2-minute walk test and the

305 modified WOMAC pain and stiffness subscales and NPRS were mild to moderate (the r

PT
306 values ranged from 0.042–0.366). This finding has significant clinical importance as it

RI
307 implies that the 2-minute walk test cannot capture some dimensions, such as pain, stiffness,

308 range of motion, and stability, which are all important dimensions of knee function.

SC
309 Therefore, performance-based tests and PROM are not interchangeable and provide distinct

310 information about different aspects of physical function after a TKA.

311
U
Moreover, we speculated that with longer-term follow-up, the scores from the
AN
312 performance-based tests might reduce due to patients getting older and tending to have more
M

313 musculoskeletal disorders or co-morbidities. In contrast, the scores from PROMs might be

314 maintained because their knees still performed well. Thus, performance-based tests and
D

315 PROM cannot replace one another, and they must therefore both be used to comprehensively
TE

316 assess the knee function of patients after TKA.

317 Our study provides new information regarding the responsiveness properties of the 2-
EP

318 minute walk test and TUG. The MCID of the 2-minute walk test and TUG 12 months after
C

319 surgery were 12.7 and 9.5, respectively. Our MCID can be used as a guide to follow each
AC

320 TKA patient routinely. If a patient tends to have a slower progression than expected from the

321 MCID, consideration should be given to interventions such as an intensive session of

322 physical therapy and more frequent follow-up and monitoring. Yuksel E et al. [38] reported a

323 minimal detectable change (MDC) score of approximately 15 meters for the 2-minute walk

324 test, which is similar to our findings. However, the data for their MDC calculations were

325 obtained on the same postoperative follow-up day after the TKA. Therefore, we believe that
15
ACCEPTED MANUSCRIPT
326 our MCID represents the true responsiveness property of the 2-minute walk test at 1 year

327 after surgery.

328 We recommend using the 2-minute walk test instead of a longer walk test, such as the

329 6-minute walk test, after a TKA. Although the 6-minute walk test has been used more often

330 in previous literature and clinical research [39-41], it might overburden patients with knee

PT
331 OA during the preoperative and early postoperative visits, and may be more difficult to

RI
332 perform routinely in clinical practice. Thus, a shorter walk test is both more clinically

333 practical and more appropriate for TKA patients. In addition, we found that the 2-minute

SC
334 walk test was well received by patients, who were most willing and able to complete the test

335 during the preoperative visit and the postoperative follow-ups. In other words, the 2-minute

336
U
walk test can be easily used to assess and monitor physical function in a clinical setting.
AN
337 Our study has several limitations. Firstly, our follow-up time of approximately 12
M

338 months might be considered short. Nevertheless, previous investigators have established that

339 TKA patients reach their plateau phase of recovery within 12 months postoperatively [28].
D

340 We therefore do not believe that a longer follow-up period would significantly change the
TE

341 results or the conclusions of our study. Secondly, this study investigated the association

342 between scores from PROMs and performance-based tests in patients who had undergone
EP

343 primary TKA. Hence, it cannot be inferred that the results would be the same for a different
C

344 population, such as patients with fractures or complex deformities. Thirdly, we chose the 2-
AC

345 minute walk test and TUG because they are simple and easy to use in routine clinical

346 practice. Nonetheless, other performance-based tests are available, and we are not able to

347 recommend which ones are best to evaluate functional recovery in patients undergoing a

348 TKA.

349 In conclusion, the association between the scores from most PROMs and the 2-minute

350 walk test is mild to moderate. Consequently, the 2-minute walk test provides distinct
16
ACCEPTED MANUSCRIPT
351 information regarding aspects of the knee function that are not captured from PROM; the 2-

352 minute walk test and PROM are therefore not interchangeable. This means that a

353 comprehensive evaluation should include a combination of PROM and performance-based

354 tests to better assess functional recovery in patients after a TKA. Performance-based tests are

355 necessary to fully characterize the change in the physical function of patients after the TKA.

PT
356 Our study showed that the 2-minute walk test is a valid performance-based test that has a

RI
357 responsiveness property with an MCID of 12.7 at 12 months after surgery. Given that the 2-

358 minute walk test is a simple and easy-to-conduct performance-based test, it can be used

SC
359 routinely in clinical practice for evaluating functional recovery after a TKA.

360

361 Conflicts of interest


U
AN
362 The authors hereby declare that there are no personal or professional conflicts of
M

363 interest regarding any aspect of this study.

364
D

365 Funding disclosure


TE

366 This research project was supported by XXX Research Fund, Grant number

367 R015732004, Faculty of Medicine, XXX Hospital, XXX University.


EP

368
C

369 References
AC

370 1. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. Validation study of

371 WOMAC: a health status instrument for measuring clinically important patient

372 relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the

373 hip or knee. J Rheumatol 1988;15:1833-40.

374 2. Podsiadlo D, Richardson S. The timed "Up & Go": a test of basic functional mobility

375 for frail elderly persons. J Am Geriatr Soc 1991;39:142-8.


17
ACCEPTED MANUSCRIPT
376 3. Stratford PW, Kennedy DM. Performance measures were necessary to obtain a

377 complete picture of osteoarthritic patients. J Clin Epidemiol 2006;59:160-7.

378 4. Bellamy N, Campbell J, Stevens J, Pilch L, Stewart C, Mahmood Z. Validation study

379 of a computerized version of the Western Ontario and McMaster Universities VA3.0

380 Osteoarthritis Index. J Rheumatol 1997;24:2413-5.

PT
381 5. Stratford PW, Kennedy DM, Maly MR, Macintyre NJ. Quantifying self-report

RI
382 measures' overestimation of mobility scores postarthroplasty. Phys Ther

383 2010;90:1288-96.

SC
384 6. Unnanuntana A, Mait JE, Shaffer AD, Lane JM, Mancuso CA. Performance-based

385 tests and self-reported questionnaires provide distinct information for the preoperative

386
U
evaluation of total hip arthroplasty patients. J Arthroplasty 2012;27:770-5.e1.
AN
387 7. Parent E, Moffet H. Comparative responsiveness of locomotor tests and
M

388 questionnaires used to follow early recovery after total knee arthroplasty. Arch Phys

389 Med Rehabil 2002;83:70-80.


D

390 8. Gandhi R, Tsvetkov D, Davey JR, Syed KA, Mahomed NN. Relationship between
TE

391 self-reported and performance-based tests in a hip and knee joint replacement

392 population. Clin Rheumatol 2009;28:253-7.


EP

393 9. Heiberg KE, Ekeland A, Bruun-Olsen V, Mengshoel AM. Recovery and prediction of
C

394 physical functioning outcomes during the first year after total hip arthroplasty. Arch
AC

395 Phys Med Rehabil 2013;94:1352-9.

396 10. Butland RJ, Pang J, Gross ER, Woodcock AA, Geddes DM. Two-, six-, and 12-

397 minute walking tests in respiratory disease. Br Med J (Clin Res Ed) 1982;284:1607-8.

398 11. Mancuso CA, Choi TN, Westermann H, Briggs WM, Wenderoth S, Charlson ME.

399 Measuring physical activity in asthma patients: two-minute walk test, repeated chair

400 rise test, and self-reported energy expenditure. J Asthma 2007;44:333-40.


18
ACCEPTED MANUSCRIPT
401 12. Unnanuntana A, Saleh A, Nguyen JT, Sculco TP, Cornell CN, Mancuso CA, et al.

402 Low vitamin D status does not adversely affect short-term functional outcome after

403 total hip arthroplasty. J Arthroplasty 2013;28:315-22.e2.

404 13. Schmitt JS, Di Fabio RP. Reliable change and minimum important difference (MID)

405 proportions facilitated group responsiveness comparisons using individual threshold

PT
406 criteria. J Clin Epidemiol 2004;57:1008-18.

RI
407 14. Quintana JM, Escobar A, Bilbao A, Arostegui I, Lafuente I, Vidaurreta I.

408 Responsiveness and clinically important differences for the WOMAC and SF-36 after

SC
409 hip joint replacement. Osteoarthritis Cartilage 2005;13:1076-83.

410 15. Hiengkaew V, Jitaree K, Chaiyawat P. Minimal detectable changes of the Berg

411
U
Balance Scale, Fugl-Meyer Assessment Scale, Timed "Up & Go" Test, gait speeds,
AN
412 and 2-minute walk test in individuals with chronic stroke with different degrees of
M

413 ankle plantarflexor tone. Arch Phys Med Rehabil 2012;93:1201-8.

414 16. Picetti GD, 3rd, McGann WA, Welch RB. The patellofemoral joint after total knee
D

415 arthroplasty without patellar resurfacing. J Bone Joint Surg Am 1990;72:1379-82.


TE

416 17. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. Validation study of

417 WOMAC: a health status instrument for measuring clinically important patient-
EP

418 relevant outcomes following total hip or knee arthroplasty in osteoarthritis. J Orthop
C

419 Rheumatol 1988;1:95–108.


AC

420 18. Kuptniratsaikul V, Rattanachaiyanont M. Validation of a modified Thai version of the

421 Western Ontario and McMaster (WOMAC) osteoarthritis index for knee

422 osteoarthritis. Clin Rheumatol 2007;26:1641-5.

423 19. Ramkumar PN, Harris JD, Noble PC. Patient-reported outcome measures after total

424 knee arthroplasty: a systematic review. Bone Joint Res 2015;4:120-7.


19
ACCEPTED MANUSCRIPT
425 20. Murray DW, Fitzpatrick R, Rogers K, Pandit H, Beard DJ, Carr AJ, et al. The use of

426 the Oxford hip and knee scores. J Bone Joint Surg Br 2007;89:1010-4.

427 21. Clement ND, MacDonald D, Patton JT, Burnett R. Post-operative Oxford knee score

428 can be used to indicate whether patient expectations have been achieved after primary

429 total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2015;23:1578-90.

PT
430 22. Rationale of The Knee Society Clinical Rating System,

RI
431 http://www.kneesociety.org/web/pdfs/knee_society_rationale_article.pdf.

432 23. Ferraz MB, Quaresma MR, Aquino LR, Atra E, Tugwell P, Goldsmith CH.

SC
433 Reliability of pain scales in the assessment of literate and illiterate patients with

434 rheumatoid arthritis. J Rheumatol 1990;17:1022-4.

435 24.
U
Spadoni GF, Stratford PW, Solomon PE, Wishart LR. The evaluation of change in
AN
436 pain intensity: a comparison of the P4 and single-item numeric pain rating scales. J
M

437 Orthop Sports Phys Ther 2004;34:187-93.

438 25. Connelly DM, Thomas BK, Cliffe SJ, Perry WM, Smith RE. Clinical utility of the 2-
D

439 minute walk test for older adults living in long-term care. Physiother Can 2009;61:78-
TE

440 87.

441 26. Dripps RD. New classification of physical status. Anesthesiol 1963;24:111.
EP

442 27. Daabiss M. American Society of Anaesthesiologists physical status classification.


C

443 Indian J Anaesth 2011;55:111-5.


AC

444 28. Mizner RL, Petterson SC, Clements KE, Zeni JA, Jr., Irrgang JJ, Snyder-Mackler L.

445 Measuring functional improvement after total knee arthroplasty requires both

446 performance-based and patient-report assessments: a longitudinal analysis of

447 outcomes. J Arthroplasty 2011;26:728-37.


20
ACCEPTED MANUSCRIPT
448 29. Terwee CB, Bot SD, de Boer MR, van der Windt DA, Knol DL, Dekker J, et al.

449 Quality criteria were proposed for measurement properties of health status

450 questionnaires. J Clin Epidemiol 2007;60:34-42.

451 30. Chan YH. Biostatistics 104: correlational analysis. Singapore Med J 2003;44:614-9.

452 31. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ:

PT
453 Lawrence Erlbaum Associates;1988.

RI
454 32. Husted JA, Cook RJ, Farewell VT, Gladman DD. Methods for assessing

455 responsiveness: a critical review and recommendations. J Clin Epidemiol

SC
456 2000;53:459-68.

457 33. Wright A, Hannon J, Hegedus EJ, Kavchak AE. Clinimetrics corner: a closer look at

458
U
the minimal clinically important difference (MCID). J Man Manip Ther 2012;20:160-
AN
459 6.
M

460 34. Stolwijk-Swuste JM, Beelen A, Lankhorst GJ, Nollet F. SF36 physical functioning

461 scale and 2-minute walk test advocated as core qualifiers to evaluate physical
D

462 functioning in patients with late-onset sequelae of poliomyelitis. J Rehabil Med


TE

463 2008;40:387-94.

464 35. Terwee CB, van der Slikke RM, van Lummel RC, Benink RJ, Meijers WG, de Vet
EP

465 HC. Self-reported physical functioning was more influenced by pain than
C

466 performance-based physical functioning in knee-osteoarthritis patients. J Clin


AC

467 Epidemiol 2006;59:724-31.

468 36. Finch E, Brooks D, Stratford PW, Mayo NE. Physical rehabilitation outcome

469 measures: a guide to enhanced clinical decision making. 2nd ed. Hamilton, ON: BC

470 Decker;2002.

471 37. Jacobs CA, Christensen CP. Correlations between knee society function scores and

472 functional force measures. Clin Orthop Relat Res 2009;467:2414-9.


21
ACCEPTED MANUSCRIPT
473 38. Yuksel E, Kalkan S, Cekmece S, Unver B, Karatosun V. Assessing Minimal

474 Detectable Changes and Test-Retest Reliability of the Timed Up and Go Test and the

475 2-Minute Walk Test in Patients With Total Knee Arthroplasty. J Arthroplasty

476 2017;32:426-30.

477 39. Fernandes L, Storheim K, Nordsletten L, Risberg MA. Development of a therapeutic

PT
478 exercise program for patients with osteoarthritis of the hip. Phys Ther 2010;90:592-

RI
479 601.

480 40. Pagani CH, Bohle C, Potthast W, Bruggemann GP. Short-term effects of a dedicated

SC
481 knee orthosis on knee adduction moment, pain, and function in patients with

482 osteoarthritis. Arch Phys Med Rehabil 2010;91:1936-41.

483 41.
U
Kennedy DM, Stratford PW, Robarts S, Gollish JD. Using outcome measure results to
AN
484 facilitate clinical decisions the first year after total hip arthroplasty. J Orthop Sports
M

485 Phys Ther 2011;41:232-9.

486
D

487
TE

488
C EP
AC
ACCEPTED MANUSCRIPT

Acknowledgements

The authors acknowledge the assistance given by Wachirapan Narktang, MSc, and Krit

Boontanapibul, MD, with data collection. The authors also acknowledge Suthipol

Udompunthurak, MSc (Applied Statistics), for his assistance with the statistical analyses.

PT
RI
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT

Figure Legend

Figure 1. Flow chart showing the patient inclusion process

Figure 2. Mean values of each patient-reported outcome measure (PROM), and within

group p-values to compare data at baseline, 3, 6 and 12 months postoperatively. (A)

PT
modified WOMAC pain; (B) modified WOMAC stiffness; (C) modified WOMAC

RI
function; (D) Oxford Knee Score; (E) modified Knee Score (KS) and; (F) Numerical Pain

Rating Scale (NPRS). The error bars indicate standard deviation.

SC
Figure 3. Mean values of the performance-based tests and within group p-values to

compare data at baseline, 3, 6 and 12 months postoperatively. (A) Two-minute walk test

U
and; (B) Timed Up and Go test. The error bars indicate standard deviation.
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT

Table 1. Patient demographic and clinical characteristics

Clinical variables Total (N=157)

Age (years), mean±SD 69.6 ± 8.1

Female gender, n (%) 137 (87.2)

Side, n (%)

PT
- Right 59 (37.6)
- Left 74 (47.1)

RI
- Bilateral 24 (15.3)

Body mass index (kg/m2), mean±SD 26.6 ± 4.4

SC
Procedure, n (%)
- Unilateral TKA 133 (84.7)
- Bilateral TKAs
U 24 (15.3)
AN
Prosthesis type, n (%)
- PFC sigma 92 (58.6)
M

- Attune 65 (41.4)
ASA classification, n (%)
D

- I 6 (3.8)
- II 124 (79.0)
TE

- III 27 (17.2)
- IV 0 (0)
EP

- V 0 (0)
Use of gait aid preoperatively, n (%)
- None 78 (49.7)
C

- Cane 49 (31.2)
AC

- Walker 29 (18.5)
- Wheel chair 1 (0.6)
Active back pain, n (%)
- Yes 19 (12.1)
- No 138 (87.9)
Number of lower extremities pain, n (%)
- Yes 35 (22.2)
- No 122 (77.8)
ACCEPTED MANUSCRIPT

Abbreviations: N, total population; SD, standard deviation; n, number; TKA, total knee
arthroplasty; PFC Sigma, Press Fit Condylar Sigma; ASA, American Society of
Anesthesiologists

PT
RI
U SC
AN
M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT

Table 2. Baseline scores of each outcome measurement

Outcome measures Mean ± standard Median Range

deviation (interquartile

ranges)

Two-minute walk test 46.2 ± 19.6 48.0 (30.0, 60.0) 2.0 – 94.5

PT
(meters) (n=157)

RI
Modified WOMAC (n=157)

- Pain 29.9 ± 9.8 30.0 (22.0, 37.0) 6.0 – 50.0

SC
- Stiffness 11.2 ± 4.9 12.0 (8.0, 15.0) 0.0 – 20.0

- Function 93.3 ± 27.6 97.0 (73.0, 113.0) 27.0 – 148.0

Oxford Knee Score (n=107)


U
16.6 ± 5.8 15.0 (13.0, 20.5) 5.0 – 34.0
AN
Modified Knee Score 58.0 ± 6.1 58.0 (56.0, 62.0) 25.0 – 89.0
M

(n=157)

Numeric Pain Rating Scale 7.4 ± 1.69 8.0 (6.0, 8.0) 2.0 – 10.0
D

(n=157)
TE

Timed Up and Go test 23.9 ± 14.3 18.9 (14.9, 29.4) 9.3 – 95.0

(seconds) (n=157)
EP

Abbreviations: WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index


C
AC
ACCEPTED MANUSCRIPT

Table 3. Construct validity of the 2-minute walk test and the modified WOMAC, Oxford Knee

Score, modified Knee Score, Numeric Pain Rating Scale and the Timed Up and Go test at

baseline, 3, 6 and 12 months after surgery

Outcome measures Two-minute walk test (meters)

PT
Baseline 3 months after 6 months after 12 months after

surgery surgery surgery

RI
Modified WOMAC

SC
Pain (n=157)

r value -0.042 -0.205 -0.366 -0.279

U
95% CI -0.197 to 0.115 -0.350 to -0.049 -0.494 to -0.222 -0.417 to -0.127
AN
p-value 0.602 0.010 <0.001 <0.001

Stiffness (n=157)
M

r value -0.090 -0.148 -0.258 -0.209

95% CI -0.243 to 0.068 -0.297 to 0.009 -0.399 to -0.106 -0.354 to -0.053


D

p-value 0.263 0.065 0.001 0.009


TE

Function (n=157)

r value -0.167 -0.280 -0.366 -0.312


EP

95% CI -0.315 to -0.010 -0.419 to -0.129 -0.494 to -0.222 -0.448 to -0.163


C

p-value 0.037 <0.00 <0.001 <0.001


AC

Oxford Knee Score (n=107)

r value 0.309 0.523 0.542 0.622

95% CI 0.125 to 0.472 0.370 to 0.649 0.392 to 0.663 0.489 to 0.726

p-value 0.001 <0.001 <0.001 <0.001


ACCEPTED MANUSCRIPT

Outcome measures Two-minute walk test (meters)

Baseline 3 months after 6 months after 12 months after

surgery surgery surgery

Modified Knee Score

PT
(n=157)

RI
r value 0.165 0.079 0.178 0.069

95% CI 0.008 to 0.314 -0.079 to 0.233 0.022 to 0.326 -0.089 to 0.224

SC
p-value 0.039 0.326 0.026 0.391

Numeric Pain Rating Scale

(n=157)
U
AN
r value -0.180 -0.167 -0.224 -0.193
M

95% CI -0.328 to -0.024 -0.315 to -0.010 -0.368 to -0.069 -0.339 to -0.037

p-value 0.024 0.037 0.005 0.016


D

Timed Up and Go test


TE

(seconds) (n=157)

r value -0.792 -0.825 -0.831 -0.781


EP

95% CI -0.844 to -0.725 -0.869 to -0.768 -0.874 to -0.775 -0.836 to -0.712

p-value <0.001 <0.001 <0.001 <0.001


C
AC
ACCEPTED MANUSCRIPT

Table 4. Responsiveness of each outcome measurement at 12 months after TKA

Outcome measures Twelve months after surgery

Effect size Standardized response mean

Two-minute walk test 1.01 1.12

Modified WOMAC

PT
- Pain 2.64 2.35

RI
- Stiffness 1.96 1.95

- Function 2.94 2.65

SC
Oxford Knee Score 3.76 3.25

U
Modified Knee Score AN 5.19 3.61

Numeric Pain Rating Scale 3.78 3.07

Timed Up and Go test 0.58 0.68


M
D
TE
C EP
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC

You might also like