Article
EDSS Less Accurate Than Other Outcome Measures for Secondary Progressive MS
Author(s):
The EDSS inaccurately reported improvement rates while paralleling disability progression rates.
Marcus W. Koch, MD, PhD
A recent study published in Neurology investigated the reliability of outcome measures in secondary progressive multiple sclerosis (SPMS) and found that the Timed 25-Foot Walk (T25FW) and 9-Hole Peg Test (9HPT) showed less random variation and measurement error than the established Expanded Disability Status Scale (EDSS).1
The EDSS showed the highest improvement rates over time and showed the least difference between progression and improvement rates. Improvement rates increased in parallel with disability progression rates. Lower improvement rates were seen in the T25W and 9HPT, with fairly stable improvement rates over time that remained below or around 10%.
Investigator Marcus W. Koch, MD, PhD, associate professor, University of Calgary, and colleagues wrote that “we had the opportunity to investigate outcome measures beyond the EDSS and found that both the T25FW and the 9HPT had lower improvement rates, consistent with less measurement error, which would suggest that they are the more reliable outcomes. The reason for this difference between the EDSS and the T25FW and 9HPT may simply lie in the fact that both the T25FW and 9HPT are objective and quantitative interval-scaled measures while the EDSS is a graded categorical measure. This difference is also relevant for inter-rater and intrarater reliability, which is relatively low for the EDSS, but high for the T25FW and 9HPT.”
READ MORE: Ibudilast Slows Retinal Thinning in Progressive MS
Koch and colleagues analyzed trial data from the 219 patients in the placebo arm of the randomized control trial (RCT) IMPACT (International MS Secondary Progressive Avonex Controlled Trial) and the 449 patients in the placebo arm of the RCT ASCEND (A Clinical Study of the Efficacy of Natalizumab on Reducing Disability Progression in Participants With Secondary Progressive Multiple Sclerosis). They compared disability progression and improvements on the different outcome measures.
Confirmation of disability progression or improvement at 3 or 6 months decreased rates of both disability progression and improvement, with no real difference observed between rates at 3- or 6-month confirmation. T25FW and 9HPT disability progression rates steadily increased over the course of the trial, while improvement rates remained under or around 10%, while the EDSS saw improvement rates increasing in parallel with disability progression rates (FIGURE). This trend was stronger in the IMPACT dataset versus the ASCEND dataset.
FIGURE. Percentage of Patients With Disability Progression and Improvement Confirmed at 3 Months in the IMPACT Database
The ratio of patients with sustained confirmed progression, as opposed to just confirmed progression, increased throughout the trial and was generally above 50% on all outcomes. The ratio was highest for the EDSS.
“Finding better therapies for progressive MS remains a large unmet need. Clinical trials in SPMS will eventually identify effective treatment options, but such trials need to be well designed to deliver reliable results. Clinical outcome measures are the most patient relevant, and will therefore likely remain the primary outcome measures in RCTs. However, their reliability is not well researched. Further research into such outcome measures in existing original trial datasets and clinical cohorts of patients with SPMS and PPMS remains necessary to improve clinical trial design,” Koch and colleagues concluded.
Tomas Kalincik, MD, PhD, neurologist, Senior Research Fellow, Melbourne Brain Centre, Melbourne University, commented on the study in a related editorial and wrote that “the study... brings to our attention the error in measurement of disability outcomes by demonstrating an incongruence among 3 commonly used disability measures. At present, most clinical trials in progressive MS use confirmed change in EDSS score as their primary or key secondary outcomes. However, as the authors elegantly show, other, more reliable clinical outcomes are needed.”2
