Baseline data from all treatment groups (N = 2333) from the International MS Secondary Progressive Avonex Controlled Trial (IMPACT), Natalizumab Safety and Efficacy in Relapsing-Remitting Multiple Sclerosis (AFFIRM), and Safety and Efficacy of Natalizumab in Combination With Interferon beta-1a in Patients With Relapsing-Remitting Multiple Sclerosis (SENTINEL) studies were combined and analyzed using Pearson product moment correlation. A significant positive correlation (r = 0.48; P < .001) was observed between baseline walking speed and baseline SF-36 PCS scores.
Change in T25-FW walking speed was assessed from baseline to 2 years in patients in the pooled placebo group. Error bars represent SEM. Clinically meaningful worsening was defined as a decrease of 5 points or more in SF-36 PCS score from baseline to 2 years. Patients who experienced a clinically meaningful worsening in SF-36 PCS score from baseline to 2 years (n = 103) exhibited a mean decline of 21.8% in walking speed vs 5.4% in those without clinically meaningful worsening in SF-36 PCS score (n = 272).
Placebo-treated patients were grouped by incremental 5% changes in T25-FW walking speed from baseline to 2 years. Error bars represent SEM. Dotted line represents the threshold for clinically meaningful change (≥5-point decrease) in SF-36 PCS score. The SF-36 PCS score declined in patients whose walking speed worsened and improved in patients who walked more than 20% faster. A decline in walking speed of 20% to 25% or more was associated with a clinically meaningful decline in SF-36 PCS score.
eTable. Baseline demographic and clinical characteristics
Cohen JA, Krishnan AV, Goodman AD, Potts J, Wang P, Havrdova E, Polman C, Rudick RA. The Clinical Meaning of Walking Speed as Measured by the Timed 25-Foot Walk in Patients With Multiple Sclerosis. JAMA Neurol. 2014;71(11):1386–1393. doi:10.1001/jamaneurol.2014.1895
Walking impairment, a common clinical manifestation of multiple sclerosis (MS), is often measured in clinical practice and clinical trials using the Timed 25-Foot Walk (T25-FW).
To evaluate the relationship between walking speed measured by the T25-FW and the Physical Component Summary (PCS) score of the 36-Item Short Form Health Survey (SF-36) to better understand the clinical meaning of T25-FW walking speed in MS.
Design, Setting, and Participants
We retrospectively analyzed data from 3 clinical trials (Natalizumab Safety and Efficacy in Relapsing-Remitting Multiple Sclerosis [AFFIRM], Safety and Efficacy of Natalizumab in Combination With Interferon Beta-1a in Patients With Relapsing-Remitting Multiple Sclerosis [SENTINEL], and International MS Secondary Progressive Avonex Controlled Trial [IMPACT]) that included T25-FW and SF-36 scores as outcomes in patients with MS. Patients had secondary-progressive MS and an Expanded Disability Status Scale score of 3.5 to 6.5 or relapsing-remitting MS and an Expanded Disability Status Scale score of 0 to 5.0.
Main Outcomes and Measures
We used Spearman rank correlation and Pearson product moment correlation (r ) and descriptive statistics to evaluate retrospectively the relationship between the SF-36 PCS score and T25-FW walking speed at baseline and the 2-year changes from baseline.
Among all 2549 patients from the 3 trials, walking speed and SF-36 PCS score at baseline were significantly correlated (n = 2333; r = 0.48; P < .001). In placebo-treated patients at 2 years, the percentage of change from baseline in walking speed was significantly correlated with the change from baseline in SF-36 PCS score (r = 0.35; P < .001). Significant correlations between the change in SF-36 PCS scores and the percentage of change in walking speed at 2 years also were observed in groups receiving active treatment (r, 0.13-0.28; P ≤ .005). Among placebo-treated patients, 27.5% had a clinically meaningful worsening (≥5-point decrease) in SF-36 PCS scores during the 2 years. Walking speed declined by 21.8% in these patients after 2 years, but only by 5.4% in those without worsening of SF-36 PCS scores.
Conclusions and Relevance
In patients with MS, walking speed measured using the T25-FW correlated with SF-36 PCS scores such that a decline in walking speed of 20% to 25% corresponded to a clinically meaningful worsening of SF-36 PCS scores. A 20% to 25% decline in walking speed may be a clinically meaningful threshold for defining worsening using the T25-FW in MS clinical trials and for monitoring patients in clinical settings.
Multiple sclerosis (MS) is a chronic demyelinating disease of the central nervous system that often leads to accumulation of irreversible disability.1 Estimates of patients with MS who experience walking impairment as the disease progresses range from 60% to more than 90%.2 The cause of walking impairment is multifactorial, with contributions from lower-limb weakness and incoordination, spasticity, sensory loss, imbalance, and fatigue.3 Walking impairment can have a negative effect on employment status, health care burden, the ability to perform activities of daily living, and health-related quality of life (HRQOL).3,4 Moreover, as many as 50% of patients with MS reported avoidance of social activities because of concerns about their walking speed.5 Maintenance of walking ability has been reported by patients with MS as the most highly valued bodily function.6
The Timed 25-Foot Walk (T25-FW) is a short-format walking test commonly used in clinical practice and is one of the components of the Multiple Sclerosis Functional Composite (MSFC), an outcome measure frequently used in clinical trials to assess disability.7 The T25-FW is easy to administer, is useful for a range of walking disabilities, and correlates well with other measures of walking ability.8 Evaluation of changes in walking ability is important in clinical trials and clinical practice to assess disability progression. The importance of understanding the clinical meaning of change in the T25-FW and other components of the MSFC was emphasized by an expert panel convened by the National Multiple Sclerosis Society in May 2011.9 That panel recommended analysis of existing data to determine the feasibility of using the MSFC approach as a primary outcome measure for clinical trials in MS. The expert panel supported the use of the MSFC as a clinically meaningful outcome measure for MS clinical trials but recommended focusing on the component measures because the meaning of changes in composite z scores is not obvious, depends on the reference population, and is not comparable across studies. To this end, the task force recommended validating changes in the MSFC components in relation to known or interpretable changes in other assessments, particularly patient-reported outcome measures.9
A clinically meaningful change is generally considered to be a change of sufficient magnitude that it is perceived by patients as being clinically important. One way to establish whether a change in a clinician-rated outcome is clinically meaningful is to determine the relationship between patient-reported outcomes and change in the clinician-rated outcome. Several previous studies assessed the clinical importance in T25-FW walking time using distribution or anchor-based approaches comparing the T25-FW with patient- or clinician-assessed measures.10- 14 Together, the findings suggest that a change of 20% or more in T25-FW walking time may be clinically meaningful. However, 1 study15 found no relationship between a worsening of 20% or more in T25-FW walking time and worsening on the Physical Impact subscale score of the patient-reported Multiple Sclerosis Impact Scale. Walking speed is more normally distributed than walking time, and the results of 2 studies16,17 that evaluated change in T25-FW walking speed suggest that improvements of 15% to 20% may be clinically meaningful. A better understanding of what defines a clinically meaningful change in T25-FW walking speed deserves further consideration.
The Natalizumab Safety and Efficacy in Relapsing-Remitting Multiple Sclerosis (AFFIRM)18 and Safety and Efficacy of Natalizumab in Combination With Interferon Beta-1a in Patients With Relapsing-Remitting Multiple Sclerosis (SENTINEL)19 clinical trials in relapsing-remitting MS (RRMS) and the International MS Secondary Progressive Avonex Controlled Trial (IMPACT) study20 in secondary-progressive MS (SPMS) included the T25-FW and the patient-reported 36-Item Short Form Health Survey (SF-36)21 as a general health status measure of HRQOL. The objective of the present analysis was to evaluate the relationship between T25-FW walking speed and the Physical Component Summary (PCS) score of the SF-36 at baseline and changes in these measures over time to better understand the meaning of walking speed measured by the T25-FW.
Descriptions of the patients and methods for the IMPACT,20 AFFIRM,18 and SENTINEL19 trials have been published. Briefly, the IMPACT trial (N = 436) was a 2-year, multicenter, randomized, double-blind trial of intramuscular (IM) interferon beta-1a, 60 μg administered weekly (n = 217), vs placebo (n = 219) in patients with SPMS. Enrolled patients ranged from 18 to 60 years of age, had clinically definite SPMS, and had an Expanded Disability Status Scale (EDSS) score of 3.5 to 6.5. The AFFIRM trial (N = 942) was a multicenter, randomized, double-blind, 30-month trial of natalizumab, 300 mg every 4 weeks (n = 627), vs placebo (n = 315). Enrolled patients ranged from 18 to 50 years of age and had a diagnosis of RRMS, an EDSS score of 0 to 5.0, and at least 1 documented relapse in the 12 months before trial entry. The SENTINEL trial (N = 1171) was a multicenter, randomized, double-blind, placebo-controlled, 30-month trial of natalizumab, 300 mg every 4 weeks, plus IM interferon beta-1a, 30 μg once weekly (n = 589), vs placebo every 4 weeks plus IM interferon beta-1a, 30 μg once weekly (n = 582). Enrolled patients ranged from 18 to 55 years of age and had a diagnosis of RRMS, an EDSS score of 0 to 5.0, at least 1 documented relapse, and had received treatment with IM interferon beta-1a once weekly for at least 12 months before randomization.
The T25-FW, as part of the MSFC, was assessed at baseline and every 3 months in the IMPACT, AFFIRM, and SENTINEL trials. The SF-36, as part of the Multiple Sclerosis Quality of Life Inventory,22 was assessed at baseline and weeks 52 and 104 in the IMPACT trial and at baseline and weeks 24, 52, and 104 in the AFFIRM and SENTINEL trials.
For the T25-FW, patients were instructed to walk as fast as they could in a safe manner along a marked 25-foot course.23 The use of a walking aid was allowed. The time in seconds to complete each test was recorded, and the test was immediately repeated. The mean T25-FW walking time from the 2 assessments was used to determine the mean walking speed (25 feet divided by the time in seconds to complete the test).
The SF-36 is a 36-item questionnaire that includes 8 multi-item health concepts (Physical Functioning, Role-Physical, Bodily Pain, General Health, Vitality, Social Functioning, Role-Emotional, and Mental Health).21 Scores range from 0 to 100, with higher scores indicating better HRQOL. The SF-36 also has 2 summary scales, the Physical Component Summary (PCS) and the Mental Component Summary, with a mean score of 50 and an SD of 10 representing the reference score for the US general population. A 5-point change in PCS score is considered clinically meaningful.24
Mean baseline and mean change from baseline to 2 years in T25-FW walking speed and SF-36 subscale and PCS scores were calculated. For the analyses of the relationship between T25-FW walking speed and SF-36 scores at baseline, data from all treatment groups of the 3 trials were combined and analyzed using Pearson product moment (r) and Spearman rank correlations. The same analyses were performed in patients with a baseline EDSS score of less than 4.0 and an EDSS score of 4.0 or greater. For the analyses of the relationship between percentage of change from baseline to 2 years in T25-FW walking speed and mean change from baseline to 2 years in SF-36 scores, data from the placebo groups of the IMPACT and AFFIRM trials were pooled and analyzed using the Pearson product moment and Spearman rank correlations. The percentage of change from baseline to 2 years in T25-FW walking speed in the pooled placebo group was summarized in patients with and without a 5-point worsening in PCS score. In addition, the change in PCS score from baseline to 2 years was summarized for patients with the following percentage of changes in T25-FW walking speed: improvement of more than 20% and declines of at least 0%, 5%, 10%, 15%, 20%, 25%, and 30%.
A total of 2549 patients constituted the combined patient population from the treatment groups of the IMPACT, AFFIRM, and SENTINEL trials. Baseline MS clinical characteristics differed among the pooled treatment groups, reflecting the different eligibility criteria among the trials (eTable in the Supplement). Most notably, the mean and median EDSS scores were higher and the proportion of patients with relapses in the year before trial entry was lower in the pooled treatment groups that included patients from the IMPACT trial vs those that included patients from the AFFIRM and SENTINEL trials. The mean baseline walking speeds by treatment group are shown in Table 1. The median walking speed in patients with MS (4.85 [range, 0.18-11.63] ft/s) was slower and the range broader than has been observed in healthy volunteers (6.76 [range, 4.81-8.93] ft/s]) (to convert feet to meters, multiply by 0.3).25 The mean (SD) SF-36 PCS score was 41.6 (10.5) among all patients and was lowest (31.7 [8.0]) in the treatment group that included only patients from the IMPACT trial, likely reflecting the effect of more severe disability on HRQOL (Table 1). Across the treatment groups, baseline SF-36 PCS scores were less than the reference score of 50.0 for the US general population.26
Among all patients in the 3 trials, a significant positive correlation (r = 0.48; P < .001) was observed between baseline walking speed and baseline SF-36 PCS scores (Figure 1 and Table 2). Among the SF-36 subscales, the strongest correlation was between the SF-36 Physical Functioning score and walking speed (Table 2). Significant correlations also were observed between baseline walking speed and SF-36 PCS and Physical Functioning scores in patients with less disability (EDSS score, <4) and more severe disability (EDSS score, ≥4) (Table 2).
In the pooled placebo groups of the IMPACT and AFFIRM trials, the mean (SD) walking speed was 4.20 (1.87) ft/s at baseline and declined by 0.44 (0.91) ft/s at year 2. The mean (SD) SF-36 PCS score was 39.9 (10.8) at baseline and worsened by 1.0 (8.4) points at year 2 (Table 3). A significant correlation was observed between the change in SF-36 PCS score and the percentage of change in walking speed from baseline to 2 years in patients in the pooled placebo group (r = 0.35; P < .001). Significant correlations between the change in SF-36 PCS score to 2 years and percentage of change in walking speed also were observed in each active treatment group (r, 0.13-0.28; P ≤ .005).
Placebo-treated patients who experienced a clinically meaningful worsening (≥5-point decrease24 [27.5%]) in SF-36 PCS score from baseline to 2 years exhibited a mean decline of 21.8% in walking speed vs a mean decline of 5.4% in those without clinically meaningful worsening in SF-36 PCS score (Figure 2). When patients were grouped by incremental changes of 5% in walking speed, SF-36 PCS score declined in patients whose walking speed worsened and improved in patients who walked more than 20% faster (Figure 3). A decline in walking speed of 20% to 25% or more was associated with a mean decline in SF-36 PCS score of 5 points or more (Figure 3).
In this post hoc analysis of data from 2 clinical trials of patients with RRMS and 1 trial of patients with SPMS, a significant correlation was observed between walking speed and SF-36 PCS score at baseline in patients with mild and more severe disability. In placebo-treated patients, a significant correlation was observed between the change from baseline in SF-36 PCS score and the percentage of change in walking speed at 2 years. In placebo-treated patients, a decline in walking speed of 20% to 25% was associated with clinically meaningful worsening in SF-36 PCS scores during the 2-year period. Thus, the results from the present anchor-based analysis demonstrate that a decline in walking speed of 20% to 25% can be considered a clinically meaningful change.
Previous work suggested a relationship between gait impairment and declines in HRQOL in patients with MS. Increased severity of patient-reported walking and mobility problems was associated with worsening health status assessed using the EuroQol 5-Dimension Index and the Hamburg Quality of Life Questionnaire in multiple sclerosis.27,28 An indirect analysis using EDSS scores as a bridge29 showed that slower walking speed was associated with lower health utility scores in patients with MS. In addition, patient-perceived reductions in walking speed, assessed using responses to item 10 on the Multiple Sclerosis Impact Scale, were associated with reduced HRQOL as assessed using the EuroQol 5-Dimension Index and the 12-Item Short Form Health Survey.30 To our knowledge, the present report is the first direct analysis of the relationship between objectively measured walking speed on the T25-FW and HRQOL in patients with MS.
Results to date suggest that a deterioration of 20% or more in walking time on the T25-FW is clinically meaningful. Distribution-based approaches showed that the variability of repeated testing of the T25-FW walking time was generally less than 20%.11,13 In anchor-based studies, a 20% or greater increase in T25-FW walking time was associated with patient-reported walking difficulty,11 and changes (worsening or improvement) of 20% or more in T25-FW walking time were associated with rank-order changes in patient-perceived disability assessed using the Guy’s Neurological Disability Scale.10 In a retrospective study of a clinical practice population of patients with primary-progressive MS, a worsening of 20% or more in T25-FW walking time was considered to be the optimal threshold for defining disease progression.31 The findings from the current analysis add to this body of evidence and suggest that a decline of 20% or more in T25-FW walking speed represents a clinically meaningful deterioration.
The clinical significance of worsening in a clinician-derived outcome measure may differ from the clinical significance of improvement in the same measure. Several studies have examined what constitutes clinically meaningful improvements in T25-FW walking time or speed. In patients with MS treated with intravenous corticosteroids, an improvement of 20% or more in T25-FW walking time was associated with patient-perceived improvement.14 In a separate analysis, T25-FW walking speed improved by 17% in patients considered minimally improved on the Clinician Global Impression scale after 14 weeks of treatment with prolonged-release fampridine (dalfampridine extended release in the United States and fampridine modified or sustained release in other countries).16 More recently, a post hoc analysis of clinical trials of prolonged-release fampridine using distribution and anchor-based approaches found that clinically meaningful changes on the patient-reported Multiple Sclerosis Walking Scale were associated with an improvement of more than 20% in T25-FW walking speed across all response criteria and as little as 15% on some response criteria.17 In the present analysis, placebo-treated patients whose walking speed improved by more than 20% had a mean improvement of 3.7 points in SF-36 PCS score during 2 years (Figure 3). This improvement did not meet the 5-point threshold for a clinically meaningful change in PCS score. However, the number of patients in this group was small (n = 29), which limited our ability to draw conclusions from these results. The magnitude of change that is clinically meaningful on the T25-FW may differ depending on whether the change represents improvement or worsening.
The present study was a post hoc analysis and thus has limitations. The results reflect those of a pooled population of patients with RRMS and SPMS from clinical trials with selective eligibility criteria and cannot be generalized across all patients with MS encountered in clinical practice. Moreover, the study design did not allow us to determine whether differences might exist with respect to thresholds for patients with RRMS and SPMS. Although most studies support a threshold of 15% to 20% for clinically meaningful change in T25-FW walking speed, the use of a different anchor might result in a different threshold value, and the thresholds for worsening and improvement in T25-FW walking speed might differ. The T25-FW measures walking over a relatively short distance and therefore does not address other attributes such as endurance or fatigue.
The present analyses demonstrated significant correlations between T25-FW walking speed and SF-36 PCS score at baseline and between declines in these measures over time. Furthermore, results using SF-36 PCS score as the anchor demonstrated that deterioration by 20% or more in T25-FW walking speed was associated with clinically meaningful worsening in patient-reported physical health status. Correlations between longitudinal changes in T25-FW walking speed and SF-36 PCS scores were weaker, but directionally the same, in the IM interferon beta-1a or natalizumab treatment groups vs the pooled placebo group. This finding might reflect in part the slower decline with treatment, which in turn would decrease the signal-to-noise ratio and thereby reduce the strength of the correlation. However, substantial worsening on the T25-FW in patients treated with IM interferon beta-1a remained, and therefore other items in the SF-36 PCS may be affected differentially by treatment with this agent. The reason for the differences in the magnitude of the correlations among the treatment groups deserves further exploration.
The more normal distribution of walking speed compared with walking time raises the question of whether T25-FW walking time in the MSFC should be transformed into walking speed before calculating z scores. The advisability of this approach deserves further consideration. Additional studies are needed to confirm the thresholds for clinically meaningful improvement and worsening in T25-FW walking speed in patients with RRMS and SPMS and to evaluate the effects of treatment and MS disease variables (eg, EDSS progression, relapse status) on the relationship between walking speed and HRQOL. In addition, a natural extension of this study would be to evaluate the relationship between SF-36 PCS score and the 9-hole peg test,32 a component of the MSFC that assesses upper limb function.
In patients with MS, walking speed measured using the T25-FW correlated with the SF-36 PCS scores such that a decline in walking speed of 20% to 25% corresponded to a clinically meaningful worsening of SF-36 PCS scores. A 20% to 25% decline in walking speed may be a clinically meaningful threshold for defining worsening using the T25-FW in MS clinical trials and for monitoring patients in clinical settings.
Accepted for Publication: June 2, 2014.
Corresponding Author: Jeffrey A. Cohen, MD, Mellen Center for Multiple Sclerosis Treatment and Research, Neurological Institute, Cleveland Clinic, U10, 9500 Euclid Ave, Cleveland, OH 44195 (email@example.com).
Published Online: September 1, 2014. doi:10.1001/jamaneurol.2014.1895.
Author Contributions: Dr Cohen had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. As corresponding author, Dr Cohen had the final responsibility for the decision to submit the paper for publication. The authors had full editorial control of the paper and provided their final approval of all content.
Study concept and design: Cohen, Krishnan, Goodman, Wang, Polman, Rudick.
Acquisition, analysis, or interpretation of data: Potts, Wang, Havrdova, Rudick.
Drafting of the manuscript: Cohen, Krishnan, Wang, Polman, Rudick.
Critical revision of the manuscript for important intellectual content: Cohen, Krishnan, Goodman, Potts, Wang, Havrdova, Rudick.
Statistical analysis: Potts, Wang.
Obtained funding: Polman.
Administrative, technical, or material support: Havrdova.
Study supervision: Rudick.
Conflict of Interest Disclosures: Dr Cohen has received personal compensation as a consultant from EMD Serono, Innate Immunotherapeutics, Genzyme, and Novartis and research support paid to his institution from Biogen Idec, Genzyme, Novartis, Receptos, Synthon, Teva, and Vaccinex. Dr Krishnan serves on the Biogen Idec Fampridine International Advisory Board and has received consulting fees from Biogen Idec and speaker fees from Bayer HealthCare, Biogen Idec, CSL Biotherapies, and Pfizer. Dr Goodman has received personal compensation for consulting services from Acorda Therapeutics, Alexion, Avanir, Biogen Idec, EMD Serono, Genzyme, Grifols, GW Pharma, Medison, Mylan, Novartis, Otsuka, Pfizer, Sanofi, Teva, and Vaccinex and financial support for research activities from Acorda Therapeutics, Biogen Idec, EMD Serono, Genzyme, Novartis, Ono, Roche, Sanofi, Sun Pharma, Takeda, and Teva. Dr Havrdova has been supported by contracts MSM 0021620849 and PRVOUK-P26/LF1/4 from the Czech Ministry of Education and has received personal compensation for consulting services and clinical trials from Biogen Idec, Merck, Novartis, Sanofi, and Teva. Dr Polman has received honoraria, consultation fees, or research support from Actelion, Bayer HealthCare, Biogen Idec, GlaxoSmithKline, Merck Serono, MorphoSys AG, Novartis, Roche, Teva, and UCB. The University of New South Wales has received research funding from Biogen Idec. No other disclosures were reported.
Funding/Support: This study was supported by Biogen Idec, including funding for editorial support in the development of this article.
Role of the Sponsor: The study sponsor designed the original clinical studies in collaboration with scientific advisory committees (Dr Cohen was the lead academic advisor for the IMPACT study; Dr Polman, for the AFFIRM study; and Dr Rudick, for the SENTINEL study). The study sponsor also participated with the authors of those studies in the collection and management of data. The study sponsor participated with the authors in the design of the current post hoc analyses, contributed to the analysis and interpretation of data, provided assistance in manuscript preparation, and reviewed and provided feedback on the manuscript to the authors.
Additional Contributions: The late Christian Confavreux, Service de Neurologie A and Fondation Eugène Devic EDMUS, Hospices Civils de Lyon, Lyon, France, contributed to the development of this study. Biogen Idec provided funding for editorial support in the development of this report; Alison Gagnon, PhD, Excel Scientific Solutions, wrote the first draft of the manuscript based on input from authors, and Elizabeth Wassmer, MLS, Excel Scientific Solutions, copyedited and styled the manuscript per journal requirements.