Cross-validation of oxygen uptake prediction during walking in ambulatory persons with multiple sclerosis

Stamatis Agiovlasitis, Robert W. Motl

Research output: Contribution to journalArticlepeer-review


BACKGROUND: An equation for predicting the gross oxygen uptake (gross-VO2) during walking for persons with multiple sclerosis (MS) has been developed. Predictors included walking speed and total score from the 12-Item Multiple Sclerosis Walking Scale (MSWS-12). OBJECTIVE: This study examined the validity of this prediction equation in another sample of persons with MS. METHODS: Participants were 18 persons with MS with limited mobility problems (42±13 years; 14 women). Participants completed the MSWS-12. Gross-VO2 was measured with open-circuit spirometry during treadmill walking at 2.0, 3.0, and 4.0 mph (0.89, 1.34, and 1.79 ms-1). REULTS: Absolute percent error was small: 8.3±6.1%, 8.0±5.6%, and 12.2±9.0% at 2.0, 3.0, and 4.0 mph, respectively. Actual gross-VO2 did not differ significantly from predicted gross-VO2 at 2.0 and 3.0 mph, but was significantly higher than predicted gross-VO2 at 4.0 mph (p < 0.001). Bland-Altman plots indicated nearly zero mean difference between actual and predicted gross-VO2 with modest 95% confidence intervals at 2.0 and 3.0 mph, but there was some underestimation at 4.0 mph. CONCLUSIONS: Speed and MSWS-12 score provide valid prediction of gross-VO2 during treadmill walking at slow and moderate speeds in ambulatory persons with MS. However, there is a possibility of small underestimation for walking at 4.0 mph.

Original languageEnglish (US)
Pages (from-to)191-197
Number of pages7
Issue number2
StatePublished - Mar 19 2016


  • Energy expenditure
  • MSWS-12
  • disability
  • gait

ASJC Scopus subject areas

  • Physical Therapy, Sports Therapy and Rehabilitation
  • Rehabilitation
  • Clinical Neurology


Dive into the research topics of 'Cross-validation of oxygen uptake prediction during walking in ambulatory persons with multiple sclerosis'. Together they form a unique fingerprint.

Cite this