There has never been an MRI study of tunnel widening comparing bioabsorbable to metal screw fixation in autologous hamstring anterior cruciate ligament (ACL) reconstruction. We randomized 62 patients to hamstring ACL reconstruction with either a bioabsorbable (n=31) or metal screw (n=31) fixation. The evaluation methods were clinical examination, KT-1000 arthrometric measurement, the International Knee Documentation Committee and Lysholm scores, and MRI. There were no differences between the groups preoperatively. Fifty-five patients (89%) were available at a minimum of 2-year follow-up (range 24-36 months). There was tunnel widening in both groups, but the increase was significantly greater in the AP dimension of the femoral tunnel in the bioabsorbable screw group compared to metal group (P=0.01). The tibial tunnels showed no intergroup difference. Ninety-four percent of the knees were normal or nearly normal according to the IKDC scores and the average Lysholm score was 91 with no intergroup difference. The follow-up AP tibial tunnel diameter was smaller with normal knee laxity compared to abnormal knee laxity. The graft failure rate in the bioabsorbable screw group was 23% (7/31 patients) and 6% (2/31 patients) in the metal screw group. The use of bioabsorbable screws resulted in more femoral tunnel widening, and more graft failures compared to metal screws. The tunnel widening in the tibia was associated with the knee laxity (P=0.02).