To compare the biomechanical characteristics, failure mode, and effects of side (left or right limb) and end (forelimb or hindlimb) of different screws in 2-screw, parallel-screw proximal interphalangeal joint arthrodesis constructs in horses. In vitro experimental study. Twenty limbs from 6 cadavers (4 complete limb sets, 2 partial sets-total of 4 forelimb and 6 hindlimb pairs). Two parallel 5.5 mm cortical (AO) screws were inserted in lag fashion in 1 randomly allocated limb of a pair, and 2 parallel headless, tapered, variable-pitched, titanium compression screws (Acutrak-Plus) were inserted in the contralateral limb. Constructs were tested in 3-point bending in a dorsopalmar (plantar) direction using a materials-testing machine at a loading rate of 5.83 mm/s. Maximal bending moment at failure and composite stiffness were calculated from data generated on load-displacement curves. Data were analyzed using a Friedman 2-way analysis of variance and Wilcoxon's signed-rank tests. No significant difference was detected for bending moment or stiffness values in proximal interphalangeal joint arthrodesis constructs using 2 parallel Acutrak-Plus or AO screws for fixation. Mean stiffness values were significantly different between forelimb and hindlimb constructs. Performance of 2 parallel Acutrak-Plus screws was biomechanically comparable with 2 parallel AO 5.5 mm cortical screws in in vitro pastern arthrodesis constructs. Acutrak-Plus screws may provide an alternative means of fixation for proximal interphalangeal joint arthrodesis.