Background Balancing the relative advantages and disadvantages of unicompartmental knee arthroplasties (UKAs) against those for TKAs can be challenging. Survivorship is one important end point; arthroplasty registers repeatedly report inferior midterm survival rates, but longer-term data are sparse. Comparing survival directly by using arthroplasty register survival reports also may be inadequate because of differences in indications, implant designs, and patient demographics in patients having UKAs and TKAs. Questions/purposes The aims of this study were to assess the survivorship of UKA in the context of one large, northern European registry, and to compare the rates of survivorship with those of cemented TKAs performed for primary knee osteoarthritis during the same 27-year period. Methods From the Finnish Arthroplasty Register, we obtained the data for 4713 patients undergoing UKAs for primary osteoarthritis (mean age, 63.5 years; minimum followup, 0 years; mean, 6.0 years; range, 0-24 years) who had surgical revision between 1985 and 2011. From this cohort, we calculated the Kaplan-Meier survivorship for revision performed for any reason and compared it with the survivorship of 83,511 patients (mean age, 69.5 years; minimum followup 0 years; mean, 6.4 years; range, 0-27 years) with TKAs treated for primary osteoarthritis during the same period. Data were adjusted for age and sex in a comparative analysis.