61 patients with shoulder impingement syndrome were randomized to management with either thoracic spinal manipulation (TSM) or a sham-TSM technique. Patients were administered two sessions of the allocated protocol, and assessed prior to the second session, immediately after the second session, and at 3-4 day follow-up. Outcomes included pain, functional measures, scapular kinematics, and scapular muscle electromyography. For pain and functional measures, differences between groups were observed to be small and clinically nonrelevant based on preset minimum clinically important differences.