Gait measures are receiving increased attention in the evaluation of patients with knee osteoarthritis (OA). Yet, there remains a need to assess variability of gait analysis in patients with knee osteoarthritis over time and how pain affects variation in these gait parameters. The purpose of this study was to determine if important gait parameters, such as the knee adduction moment, knee flexion moment, peak vertical ground reaction force, and speed, were repeatable in patients with mild-to-moderate knee OA over a trial period of 12 weeks. Six patients were enrolled in this cross-over study design after meeting strict inclusion criteria. Gait tests were conducted three times at 4 week intervals and once after the placebo arm of a randomized treatment sequence; each gait test followed a 2-week period of receiving a placebo for a pain modifying drug. Repeatability for each gait variable was found using intraclass correlation coefficients (ICC) with a two-way random model. This study found that the knee adduction moment was repeatable throughout the four gait tests. However, normalized peak vertical ground reaction force and knee flexion moment were not as repeatable, varying with pain. This suggests that these gait outcomes could offer a more objective way to measure a patient's level of pain.