Objective: To determine speed dependence of crutch force and oxygen uptake, and to discuss the implications of differences in self-selected walking speed between orthoses in a comparative trial. Design: Cross-sectional comparison. Setting: Treadmill experiments and gait laboratory experiments were performed at five and three different imposed walking speeds, respectively. Patients: Five paraplegic subjects with lesions between T9 and T12 were included. All subjects had experience with ambulation using the advanced reciprocating gait orthosis (ARGO) as well as walking on a treadmill. Main Outcome Measures: Crutch force time integral (CFTI), crutch peak force on stance and swing side (CPFstance and CPFswing), oxygen uptake (Vo2), oxygen cost (Eo2). Results: Vo2, Eo2, and CFTI were strongly dependent on walking speed. CPFstance and CPFswing were less dependent. However, depending on the clinically relevant difference that should be detected in a comparative trial, the peak forces can still be confounded by walking speed. Conclusion: CFTI, CPFswing, Vo2, and Eo2 should be adjusted for walking speed if differences in walking speed between orthoses are found, but this correction is relevant only if there is no effect modification. Such modification (different slopes between orthoses) cannot be excluded for the studied outcome measures. In addition, because determination of effect modification is difficult in small studies, standardization of walking speed, by means of a three-point design, is recommended.