OBJECTIVE: To evaluate the effect of unicortical and bicortical screw holes on residual bone strength by comparing the in vitro torsional strength of cadaveric sheep tibiae with screw holes from plates with unicortical and bicortical purchase relative to each other and to intact bone. DESIGN: The paired tibiae were grouped randomly and torsion tested to failure as follows: Group I -- unicortical screw holes versus intact bone; Group II -- bicortical screw holes versus intact bone; and Group III -- bicortical versus unicortical screw holes. BACKGROUND: Recently the point contact fixator, or PC-Fix (using screws with unicortical purchase), was designed to minimize bone devascularization beneath the plate compared to the conventional dynamic compression plate, or DCP (bicortical purchase), and possibly reduce refracture rates after plate removal. However, the effects of unicortical versus bicortical residual screw holes on potential refracture are unknown. METHODS: Screw holes were drilled and tapped through either a 7-hole bicortical DCP or a unicortical PC-Fix in 18 paired cadaveric sheep tibiae. The bones were then tested in torsion. RESULTS: The average decreases in the torsional strength of unicortical screw holes versus intact bone, bicortical screw holes versus intact bone, and bicortical versus unicortical screw holes were 21.6, 31.4, and 26.7% respectively. Mean torque values to failure were statistically significant within each of the three groups: P<0.01, 0.001, and 0.001 respectively. CONCLUSIONS: Bones with unicortical screw holes are significantly weaker in torsion than intact bones, but significantly stronger than bones with bicortical holes. RELEVANCE: Refracture after hardware removal is a recognized clinical complication of the traditional bicortical screw fixation. In combination with the demonstrated biological advantages, implants with unicortical purchase may be associated with less risk for early refracture upon removal.