The aim of the present study was to evaluate the importance and the necessity of metabolic measurements to quantify locomotor impairment in a clinical context. Oxygen consumption, heart rate, pulmonary ventilation and walking speed were measured during locomotion in 14 normal subjects, used as a control group, and 82 patients with different pathologies [hemiparetic, paraparetic, tetraparetic, orthopaedic and paraplegic patients, who walked using a reciprocating gait orthosis (RGO)]. The subjects were characterized on the basis of a cumulative impairment score (CIS), based on clinical scales commonly used to evaluate impairment and disability in locomotion. Appropriate indices of energy, cardiac and ventilatory costs expressed per metre walked, globally called physiological costs, were obtained. It resulted that the most comfortable speed (MCS) of normal subjects was significantly higher than that of each group of patients. Normal subjects' physiological costs were found to be significantly lower than those of patients who needed either a device or the help of a person to walk. All measured parameters correlated significantly with each other. The MCS was found to be the most correlated parameter with the CIS (r = 0.8), and therefore it must be considered the best single measurement, if only one is to be used. Measurements more precise than MCS, such as the physiological costs, may be necessary in clinical trials.