Background: Unicompartmental knee arthroplasty (UKA) is performed as an alternative to total knee arthroplasty (TKA) and high tibial osteotomy for unicompartmental osteoarthritis. Objectives: We examined whether the tolerable range of component malalignment is narrower in obese (BMI > 30) or in nonobese patients. Methods: We performed 163 consecutive all-poly medial UKA from 01/01/1995 to 31/10/2003. We examined 83 patients (88 knees) with a minimal follow-up period of 10 years. We examined the correlation between limb- and component malalignment and clinical outcome separately in the obese (67 knees) and nonobese (21 knees) groups. Results: The 10-year prothesis survival was 92.8%, and 9 UKA were converted to total knee arthroplasty. The average time for revision was 84.44 (48 to 144) months. The 8 obese and 1 nonobese patients had slightly higher BMI (33.47) than the 83 long-term followed patients (31.72). In each of these 9 patients, knee score and functional score were poor. At every revision, we used stems and augments. In the obese group, the prosthetic joint space depression correlated with fair and poor knee and functional scores, the prosthetic joint space elevation correlated with degenerative changes in the lateral tibiofemoral joint. Conclusions: In the obese group, we noted at least 2 mm of prosthetic joint space depression in all of the 8 failed knees, and 4 mm or more than 4 mm in 6 cases. We hypothesize that the reason fot the subsidence of the tibial component is the increased loading because of prosthetic joint space depression. The result of this study suggests that tibial component positioning which provides an optimal level of prosthetic joint space reduces the risk of failure in medial UKA, prevents degenerative changes in lateral tibiofemoral joint, and provides better long-term clinical outcome.