36 patients scheduled for total shoulder arthroplasty were randomized to shoulder immobilization for 6 weeks in either neutral rotation or internal rotation. Patients were assessed for patient-reported outcome measures and shoulder range of motion over the first postoperative year. For patient-reported function, no significant differences were observed between groups for the Disabilities of the Arm, Shoulder, and Hand (DASH) score, the Western Ontario Osteoarthritis of the Shoulder (WOOS) score, or the Singe Assessment Numeric Evaluation (SANE) score. For range of motion, a significant difference in favour of the neutral rotation group compared to the internal rotation group was observed in the increase in abducted external rotation while supine. Additionally, decrease in pain during shoulder adduction was significantly greater in the neutral rotation group compared to the internal rotation group.