Background: Superior capsule reconstruction (SCR) was developed to improve shoulder function and alleviate pain in patients with primary irreparable rotator cuff tears. Although suitable clinical results of SCR have been reported, only a few studies have investigated the clinical outcomes of arthroscopic superior capsule reconstruction (ASCR) using tensor fascia at a minimum follow-up period of 2 years after surgery. Methods: Among 100 consecutive patients who underwent ASCR since April 2010, this retrospective cohort study included 49 patients with a Hamada classification of <= 3 who were available for at least 2-year follow-up after surgery. The mean follow-up period was 34.5 (24-74) months. We analyzed preoperative and postoperative Japanese Orthopaedic Association (JOA) scores, University of California at Los Angeles (UCLA) shoulder scores, preoperative and postoperative active elevation angles, external and internal rotations with the arms in the anatomical position, manual muscle test (MMT) scores, preoperative and postoperative acromiohumeral distance (AHD), and cuff integrity on postoperative magnetic resonance imaging using the Hasegawa classification. We compared 27 pseudoparalyzed (elevation of <90(degrees)) shoulders with 22 nonpseudoparalyzed shoulders. We also evaluated the treatment of patients with subscapularis tendon tears and compared the outcomes of patients with good graft repair and those with graft retear. Results: The JOA score, UCLA score, active elevation angle, internal and external rotation angles, and muscle strength (MMT) significantly improved at the time of investigation preoperatively and 2 years postoperatively. The mean AHD also improved from 5 +/- 2.6 mm preoperatively to 9 +/- 2.8 mm postoperatively. No significant difference in graft tear rate was observed between pseudoparalyzed shoulder and nonpseudoparalyzed shoulder groups. The subscapularis tendon was torn in 26 of 49 (53%) patients, and all patients underwent repair. The graft repair group showed a significant improvement in JOA scores, UCLA shoulder scores, joint range of motion, MMT, and AHD postoperatively, but not in internal rotation strength. In contrast, the graft tear group did not show any significant improvement. All patients could return to work, except for those performing heavy labor. Complications included graft tear in five patients, postoperative infection in two patients, and progressive postoperative arthropathic changes in one patient. Conclusion: Good clinical results of ASCR were obtained using tensor fascia lata at 2 years after surgery, with few complications and low graft tear rates.