BACKGROUND: The aim of this study is to assess clinical results of total knee arthroplasty (TKA) with pre-operative patella alta or patella baja, compared to TKA with pre-operative normal patellar height. METHODS: Patella height was measured using the Blackburne-Peel (BP) method in an initial, prospective cohort of 4103 TKAs performed at a single center. Three groups were defined: normal patella height (0.54 < BP index ? 1.06), patella alta (BP index >1.06), and patella baja (BP index ?0.54). Pre-operative and post-operative clinical and radiological characteristics were collected and compared. All prostheses used a system of posterior stabilization by a third condyle. The primary outcome measure was the post-operative Knee Society Score. Statistical analysis was performed using Student's t-test and chi-squared test, with P < .05. RESULTS: Two hundred twenty-three TKAs with pre-operative patella alta, 307 TKAs with pre-operative patella baja, and 2248 with pre-operative normal patella height, mean follow-up 39.3 months (range 24-239), were ultimately included. Post-operatively, there was no difference between the 3 groups in terms of Knee Society Score (knee and function scores). Patella baja statistically decreased the maximal flexion (118.2° vs 115.5°, P = .002) and increased the rate of patellar fracture (0.9% vs 2%, P = .05). Finally, survival rates were similar in the 3 groups with each above 91% (±2%) at 10 years. CONCLUSION: Clinical and radiological results for TKAs with pre-operative patella alta and patella baja were comparable to TKAs with a normal pre-operative patellar height. Risk of post-operative patellar fracture increased for patients with pre-operative patella baja.