We conducted a prospective, randomized, controlled trial to compare functional outcomes, complications, and reoperation rates in elderly patients with displaced intra-articular, distal humeral fractures treated with open reduction-internal fixation (ORIF) or primary semiconstrained total elbow arthroplasty (TEA). Forty-two patients were randomized by sealed envelope. Inclusion criteria were age greater than 65 years; displaced, comminuted, intra-articular fractures of the distal humerus (Orthopaedic Trauma Association type 13C); and closed or Gustilo grade I open fractures treated within 12 hours of injury. Both ORIF and TEA were performed following a standardized protocol. The Mayo Elbow Performance Score (MEPS) and Disabilities of the Arm, Shoulder and Hand (DASH) score were determined at 6 weeks, 3 months, 6 months, 12 months, and 2 years. Complication type, duration, management, and treatment requiring reoperation were recorded. An intention-to-treat analysis and an on-treatment analysis were conducted to address patients randomized to ORIF but converted to TEA intraoperatively. Twenty-one patients were randomized to each treatment group. Two died before follow-up and were excluded from the study. Five patients randomized to ORIF were converted to TEA intraoperatively because of extensive comminution and inability to obtain fixation stable enough to allow early range of motion. This resulted in 15 patients (3 men and 12 women) with a mean age of 77 years in the ORIF group and 25 patients (2 men and 23 women) with a mean age of 78 years in the TEA group. Baseline demographics for mechanism, classification, comorbidities, fracture type, activity level, and ipsilateral injuries were similar between the 2 groups. Operative time averaged 32 minutes less in the TEA group (P = .001). Patients who underwent TEA had significantly better MEPSs at 3 months (83 vs 65, P = .01), 6 months (86 vs 68, P = .003), 12 months (88 vs 72, P = .007), and 2 years (86 vs 73, P = .015) compared with the ORIF group. Patients who underwent TEA had significantly better DASH scores at 6 weeks (43 vs 77, P = .02) and 6 months (31 vs 50, P = .01) but not at 12 months (32 vs 47, P = .1) or 2 years (34 vs 38, P = .6). The mean flexion-extension arc was 107 degrees (range, 42 degrees -145 degrees) in the TEA group and 95 degrees (range, 30 degrees -140 degrees) in the ORIF group (P = .19). Reoperation rates for TEA (3/25 [12%]) and ORIF (4/15 [27%]) were not statistically different (P = .2). TEA for the treatment of comminuted intra-articular distal humeral fractures resulted in more predictable and improved 2-year functional outcomes compared with ORIF, based on the MEPS. DASH scores were better in the TEA group in the short term but were not statistically different at 2 years' follow-up. TEA may result in decreased reoperation rates, considering that 25% of fractures randomized to ORIF were not amenable to internal fixation. TEA is a preferred alternative for ORIF in elderly patients with complex distal humeral fractures that are not amenable to stable fixation. Elderly patients have an increased baseline DASH score and appear to accommodate to objective limitations in function with time.