Patellar Tendon Versus 4-Strand Semitendinosus and Gracilis Autografts for Anterior Cruciate Ligament Reconstruction: A Meta-analysis of Randomized Controlled Trials With Mid- to Long-Term Follow-Up

Purpose: To compare patellar tendon (PT) autografts with 4-strand semitendinosus and gracilis (STG) autografts concerning knee stability, functional outcomes, as well as complications after anterior cruciate ligament (ACL) reconstruction at minimal follow-up of 5 years. Methods: A systematic search of the literature was performed in PubMed, Embase, and the Cochrane Library to identify published prospective randomized controlled trials on clinical studies comparing PT autograft and 4-strand STG autografts for ACL reconstruction. The results of the eligible studies were analyzed in terms of knee stability (laxity measurements, Lachman test, and pivot-shift test), functional outcomes (Lysholm score, International Knee Documentation Committee score, Cincinnati score, Tegner score, single-legged hop test, and return to preinjury activity level), and complications (loss of range of motion [ROM], pain, graft re-rupture, revision, and osteoarthritis [OA]). Results: Nine clinical studies with 630 patients (313 PT and 317 STG autografts) met the inclusion criteria. No statistically significant differences were found between the PT and STG group in Lachman test, pivot-shift test, International Knee Documentation Committee score, Cincinnati score, loss of ROM, kneeling pain, graft re-rupture rate, revision rate, and OA rate. The STG group was found with less anterior knee pain (P = .003). There were no clinically significant differences for the outcomes of SSD, Lysholm score, and Tegner score. Conclusions: Except for significantly greater risk of anterior knee pain, PT autograft had comparable results with 4-strand STG autograft in terms of knee stability and functional outcomes after ACL reconstruction with mid- to long-term follow-up. Besides, we found no statistically significant difference in loss of ROM, kneeling pain, graft re-rupture rate, revision rate, and OA change, but these results were underpowered.