The objective of this study was to evaluate the initial fixation strength of a metal interference screw (9 x 25 mm; Arthrex, Naples, FL) with respect to its position relative to the bone plug in anterior cruciate ligament reconstruction using a bone-patellar tendon-bone graft. Porcine lower limbs were used. In group 1 (n = 20) the screw was placed adjacent to the cortex of the bone plug. In group 2 (n = 20) the screw was placed between the cancellous surfaces of the tibial tunnel and the bone plug. Specimens were loaded to failure using an universal material testing machine. Failure mode was defined by visual analysis of the specimens after loading. Primary fixation strength was not significantly different for both groups. Maximum force was 779 +/- 179 N (group 1, cortical side) and 743 +/- 143 N (group 2, cancellous side). There was a difference in failure modes. In group 1, testing resulted predominantly in tendon ruptures and bone plug fractures (70%). In group 2, bone plug pullouts were predominant (75%). From these data we conclude that the position of the screw relative to the bone plug influences failure mode but not the initial fixation strength.