Walking is a very common activity for the human body. It is so common that the musculoskeletal and cardiovascular systems are optimized to have the minimum energetic cost at 4 km/h (spontaneous speed). A previous study showed that lumbar and thoracolumbar adolescent idiopathic scoliosis (AIS) patients exhibit a reduction of shoulder, pelvic, and hip frontal mobility during gait. A longer contraction duration of the spinal and pelvic muscles was also noted. The energetic cost (C) of walking is normally linked to the actual mechanical work muscles have to perform. This total mechanical work (W(tot)) can be divided in two parts: the work needed to move the shoulders and lower limbs relative to the center of mass of the body (COM(b)) is known as the internal work (W(int)), whereas additional work, known as external work (W(ext)), is needed to accelerate and lift up the COM(b) relative to the ground. Normally, the COM(b) goes up and down by 3 cm with every step. Pathological walking usually leads to an increase in W (tot) (often because of increased vertical displacement of the COM(b)), and consequently, it increases the energetic cost. The goal of this study is to investigate the effects of scoliosis and scoliosis severity on the mechanical work and energetic cost of walking. Fifty-four female subjects aged 12 to 17 were used in this study. Thirteen healthy girls were in the control group, 12 were in scoliosis group 1 (Cobb angle [Cb] < or = 20 degrees), 13 were in scoliosis group 2 (20 degrees < Cb < 40 degrees), and 16 were in scoliosis group 3 (Cb > or = 40 degrees). They were assessed by physical examination and gait analysis. The 41 scoliotic patients had an untreated progressive left thoracolumbar or lumbar AIS. During gait analysis, the subject was asked to walk on a treadmill at 4 km h(-1). Movements of the limbs were followed by six infrared cameras, which tracked markers fixed on the body. W(int) was calculated from the kinematics. The movements of the COM(b) were derived from the ground reaction forces, and W(ext) was calculated from the force signal. W(tot) was equal to W(int) + W(ext). Oxygen consumption VO2 was measured with a mask to calculate energetic cost (C) and muscular efficiency (W(tot)/C). Statistical comparisons between the groups were performed using an analysis of variance (ANOVA). The external work (W(ext)) and internal work (W(int)) were both reduced from 7 to 22% as a function of the severity of the scoliosis curve. Overall, the total muscular mechanical work (W(tot)) was reduced from 7% to 13% in the scoliosis patients. Within scoliosis groups, the W(ext) for the group 1 (Cb > or = 20 degrees) and 2 (20 < or = Cb < or = 40 degrees) was significantly different from group 3 (Cb > or = 40 degrees). No significant differences were observed between scoliosis groups for the W(int). The W(tot) did not showed any significant difference between scoliosis groups except between group 1 and 3. The energy cost and VO2 were increased by around 30%. As a result Muscle efficiency was significantly decreased by 23% to 32%, but no significant differences related to the severity of the scoliosis were noted. This study shows that scoliosis patients have inefficient muscles during walking. Muscle efficiency was so severely decreased that it could be used as a diagnostic tool, since every scoliosis patient had an average muscle efficiency below 27%, whereas every control had an average muscle efficiency above 27%. The reduction of mechanical work found in scoliotic patients has never been observed in any pathological gait, but it is interpreted as a long term adaptation to economize energy and face poor muscle efficiency. With a relatively stiff gait, scoliosis patients also limit vertical movement of the COM(b) (smoothing the gait) and consequently, reduce W(ext) and W(int). Inefficiency of scoliosis muscles was obvious even in mild scoliosis (group 1, Cb < 20 degrees) and could be related to the prolonged muscle contraction time observed in a previous study (muscle co-contraction).