BACKGROUND: The first metatarsophalangeal joint may be fused in order to treat arthritis or instability. The use of shape-memory staples for fixation is well recognised, but little work has been done into the optimal configuration of staples. METHODS: The structural behaviour of first metatarsophalangeal joint (MTPJ) arthrodeses using shape-memory staples or crossed screws was studied using cadaveric porcine joints. Five fixation configurations were tested: single vertical or horizontal staple, paired staples in dorsal-medial configuration (0-90° to the sagittal plane), paired staples in oblique orthogonal configuration (45-135°); or two crossed screws. Specimens were loaded in cyclical dorsiflexion for 1000 cycles. Plantar gapping and shearing were measured. Specimens were then loaded to failure. RESULTS: Cyclic testing caused more shear in the 45-135° staples than the crossed screws (1.0mm±0.5mm compared to 0.14mm±0.4mm, p<0.01). No significant difference was found in plantar gap formation. Single vertical and horizontal single staples failed at 15N and 19. CONCLUSIONS: N, respectively. Paired 0-90° staples failed at 43N±9N, significantly lower than the 45-135° staples (141N±25N; p<0.001) and crossed screws (180N±67N; p<0.001). There was no significant difference between the 45-135° staples and crossed screws. Screws failed by sudden cortical fracture; staples displayed gradual pull-out and shearing. First MTPJ arthrodeses fixed with single staples are not recommended. Arthrodeses fixed with staples at 0-90° to the sagittal plane were significantly less strong than two crossed screws. However, positioning oblique staples at 45-135° significantly improved stability, creating a construct as strong as, crossed screws. None of the constructs was strong enough for immediate weight bearing.