The energy cost of walking is a marker of metabolic solicitation and allows energy quantification of walking. It is strongly correlated to the patient's walking speed with post-stroke sequelae. We wanted to develop a model for predicting the energy cost of walking to create of walking distance recommendations aimed at achieving recommended energy expenditure levels to ensure health benefits (> 1000kcal/week during physical activities).

To verify the close relationship between spontaneous walking speed (S_free) and energy cost of walking at S_free (Cw_free) in subjects with post-stroke hemiparetic patients sequelea and to test the validity of a prediction model to estimate Cw_free based on S_free.

Twenty-six mild to moderately disabled chronic participants with stroke aged 65 yrs (± 15.7yrs) walked at S_free using mobility aids if necessary, for 6minutes. The Cw_free was measured at a stabilized metabolic rate by indirect calorimetry using the Metamax3B device. The relationship between S_free and Cw_free was analyzed using the correlation coefficient, coefficient of determination. The Cw_free prediction model was developed from a regression equation and then tested on a new population with the same inclusion/exclusion criteria (n=29).

S_free and Cw_free had a correlation coefficient of −0.94 and a R² of 0.97 enabling to formulate a regression equation and develop the Cw_free prediction model based on S_free. The prediction model tests yielded accurate results with a mean bias of −0.02mL.kg⁻¹.m⁻¹–MD%=4%. The limits of agreement were −0.31; 0.26mL.kg⁻¹.m⁻¹ MD%=30.9% of the mean measured Cw_free.

Cw_free can be estimated precisely by a simple measurement of the walking speed. In this way, we can easily define the distance the subject must travel according to his Cw to reach the recommended energy expenditure targets. It is a great help for the practitioner for recommendations in the field of physical activity after stroke.