17.10.2026 · 6 min read
Two people can walk at the same speed but spend a different amount of energy on each metre. The metabolic cost of walking is expressed, for example, in joules per kilogram of body mass per metre or in millilitres of oxygen per kilogram per metre. The lower the cost under comparable conditions, the more economical the movement.
VO₂max answers a different question: what is the maximal capacity of the body to use oxygen during intense exercise. Training can raise VO₂max, lower the cost of walking, or do both. In either of the latter cases an ordinary pace takes up a smaller share of the available aerobic power and may subjectively feel easier.
The cost depends strongly on speed. Its relationship with speed is U-shaped: walking that is too slow or too fast usually costs more per unit of distance than an individually comfortable range. Tests at different speeds therefore cannot be compared directly.
With age, step length and cadence change, as do the work of the ankle and hip joints, stability, coordination and the co-contraction of antagonist muscles. The systematic review by Das Gupta et al. showed that in older adults the net cost of walking was on average 17% higher than in young adults at a similar average speed.
After stroke, or in Parkinson's disease or multiple sclerosis, additional roles may be played by impaired motor control, asymmetry, spasticity and excessive muscle activation. As a result, a slow walk can require a large share of maximal power and quickly cause fatigue.
Hortobágyi and colleagues searched for randomised studies of chronic training in adults without neurological disease and in people with stroke, Parkinson's disease or multiple sclerosis. The analysis included 27 studies, 36 training groups, 54 outcomes and 1,177 participants: 703 relatively healthy adults and 474 participants with neurological conditions.
Sixteen studies were conducted in healthy adults, nine after stroke, and one each in Parkinson's disease and multiple sclerosis. The median programme lasted 12 weeks, included 36 sessions and was usually held three times a week. Aerobic, strength, combined and other interventions were studied.
Overall, training reduced the metabolic cost of walking: g = 0.40, 95% confidence interval 0.24–0.56, I² about 46%. This is a standardised effect that pools different units and protocols. It cannot be converted into a universal number of kilocalories saved.
In healthy participants the average effect was g = 0.49. In people after stroke the estimate was smaller — g = 0.24 — and did not differ statistically from zero at p = 0.065. The single studies in Parkinson's disease and multiple sclerosis are too few for general recommendations for these groups.
The type, intensity, frequency and duration of training were not significant moderators. But the authors specifically warn that the meta-regression may not have had the power to detect real differences.
Aerobic training can improve oxygen delivery and use, vascular function and mitochondria. Strength training can increase the available force and reduce the relative effort of each step. Tasks involving walking, balance and coordination can improve the timing of muscle activity and reduce unnecessary co-contraction.
These mechanisms are plausible, but the meta-analysis could not determine which of them is responsible for the overall effect. Improvements in the cost of walking had no convincing relationship with changes in body composition, mobility, strength, power or balance; the relationship with metabolic outcomes was only a trend.
In the post-stroke subgroup no statistically convincing average effect was obtained. This is not proof that rehabilitation is of no benefit. Training can improve speed, endurance, independence and cardiovascular fitness even if the cost of each metre changes little.
After stroke, the causes of an uneconomical gait are varied, and the ability to change an established pattern is limited by the severity and location of the lesion. The programmes and the baseline condition of participants also differed. Clinical rehabilitation therefore cannot be chosen on the basis of a single summary indicator from a meta-analysis.
Nine studies had a low risk of bias, 11 had some concerns, and seven had a high risk. Outcome measurement was especially vulnerable: repeated familiarisation with the treadmill or the test could itself improve economy.
Moderate heterogeneity means that the effect was not the same everywhere. There are few data for Parkinson's disease and multiple sclerosis. The authors cannot explain the exact mechanism or establish the best programme. The composition of the samples also does not allow the findings to be freely transferred to young athletes, who already have an economical gait.
For a healthy adult, regular aerobic and strength training remains a sensible strategy for maintaining function. But most people do not need to deliberately "optimise the cost of walking": a programme that can be performed regularly and safely matters more.
For an older person, a combination of walking, strength and balance can reduce the relative difficulty of everyday movement. People with neurological conditions need an individual programme from a physiotherapist or doctor: average data do not determine the approach for a specific patient.
If walking suddenly becomes much harder, or chest pain, marked breathlessness, weakness or a new neurological symptom appears, this is not a question of sport economy and requires medical assessment.
Chronic training on average makes walking more energy-efficient, especially in relatively healthy adults. However, "walking feels easier" can mean several adaptations: a lower cost per metre, a higher VO₂max, greater strength or better coordination. The new review confirms the overall effect but does not prescribe a single best type of training.
Tags: rehabilitation, walking, exercise physiology, ageing, metabolic cost
Research on training, nutrition and recovery: the key points worth knowing. One email a week with the new Sport in Luxembourg articles: title, short description and a link to read.
We use cookies for visitor analytics. You can accept or decline their use. Learn more in our Privacy Policy
