The natural control of prosthetic robotic hands via surface electromyography (sEMG) remains a challenge despite the flexor-extensor muscular system of the fingers is usually partially preserved in patients with trans-radial amputations. In this work we analyze the Ninapro database (Non Invasive Adaptive Hand Prosthetics, which is currently the largest sEMG database of hand movements. The aim of the work is to identify relationships between clinical parameters of the amputation and movement recognition accuracy, in order to foster the integration between amputation surgery and innovative robotic hand prostheses.