Previous in vivo studies on total ankle arthroplasty (TAA) kinematics were mainly performed using skin marker analysis, which has the drawback of skin movement artefacts 1 . A further limitation is the inaccessibility of the talus for attaching markers, thus the impossibility to distinguish tibiotalar from subtalar motion. So far it is not known how well skin marker analysis detects the kinematics of the TAA.

The kinematics of 11 TAA participants were simultaneously analysed by skin marker and videofluoroscopic assessment during level gait (gt), walking up- (uph) and downhill (dnh). The fluoroscopic data analysis included a 2D/3D registration (error < 0.2° in-plane, <1.3° out-of-plane) 2 . The markerset consisted of 4 rearfoot and 6 shank markers 3 . For both approaches joint rotations were described along the axes of the marker based joint coordinate system. As a descriptor of differentiation the maximal and the root mean square differences (max diff, RMS diff) between skin marker and fluoroscopic joint rotations were calculated over the whole stance phase. Besides, maximal ranges of motion (ROM) were compared using a paired t-test.

Skin marker analysis significantly overestimated sagittal plane ROM of the TAA for 5(gt), 6(uph) and 6(dnh) and underestimated for 1(uph) and 2(dnh) subjects. Frontal plane ROM was significantly overestimated for 7(gt), 8(uph) and 9(dnh) of the 11 subjects. Transverse plane ROM was for 2(uph) and 2(dnh) subjects significantly overestimated, and for 3(gt), 1(uph) and 7(dnh) subjects significantly underestimated by skin markers. For mean RMS diff, mean max diff and mean ROM see Table 1.

The differences between skin marker assessed rearfoot-shank and the fluoroscopic assessed isolated TAA motion were neither consistent between subjects, nor motion planes, nor conditions. For transverse and frontal plane rotations, the maximal differences were in the range of the maximal corresponding ROM. Discrepancies for the sagittal plane were smaller, but still for some subjects, ROM were significantly different.