The object of this prospective study is to compare the Insall-Salvati ratio between the patients who have an anterior cruciate ligament (ACL) tear and receive arthroscopic-assistant ACL reconstruction and the patients who have no ACL tear but do have an internal disorder of the knee and receive arthroscopic surgery. We prospectively and consecutively collected into two groups a total of 217 patients who had sport injuries and received arthroscopic surgery. The study group included 115 patients who had an ACL tear and received arthroscopic-assistant ACL reconstruction with middle-third bone-patella tendon-bone graft. The control group included 102 patients with internal disorders of the knee joint, including meniscus tear, plicae, or other chondral lesion, but without an ACL tear. We measured the patellar Insall-Salvati ratio [12] on the pre-operative X-ray films for all patients. The Insall-Salvati ratio in the ACL-tear study group is significantly smaller than the control group of internal disorders of the knee (0.99+/-0.11 vs 1.05+/-0.12, p=0.001). There is no significant difference in age, gender, the side of the involved knee, duration of symptoms, patella length and patella tendon length between the two groups. In conclusion, our study shows that patella infra has an association with ACL tears, and patella infra may be a risk factor for ACL tears. In patients with an ACL tear who had patella baja, the middle-third patellar tendon may not be an ideal graft for reconstruction.