The investigators reviewed 245 fixed-bearing unicondylar arthroplasties that one surgeon performed as treatment of medial compartment osteoarthritis between 1988 and 1997 using a variety of cemented metal-backed tibial components and gamma-irradiated-in-air polyethylene bearings. Multivariate statistical analysis was used to evaluate how the event of revision was influenced by 3 patient factors, 3 implant factors, and 7 factors assessed from preoperative and early postoperative radiographs. Five factors were statistically associated with revision: (younger) patient age, (thinner) tibial component initial thickness, (longer) polyethylene shelf age, (lesser) angular reduction of medial tibial plateau varus, and (more varus) postoperative hip-knee-ankle angle. Besides illustrating deleterious consequences of using gamma-irradiated-in-air polyethylene in medial unicompartmental arthroplasty, our results support reducing varus angulation of the medial tibial plateau and knee at surgery.