Background and purpose There is no substantial clinical evidence for the superiority of alternative bearings in total knee arthroplasty (TKA). We compared the short-term revision risk in alternative surface bearing knees (oxidized zirconium (OZ) femoral implants or highly crosslinked polyethylene (HXLPE) inserts) with that for traditional bearings (cobalt-chromium (CoCR) on conventional polyethelene (CPE)). The risk of revision with commercially available HXLPE inserts was also evaluated.Methods All 62,177 primary TKA cases registered in a Total Joint Replacement Registry between April 2001 and December 2010 were retrospectively analyzed. The endpoints for the analysis were all-cause revisions, septic revisions, or aseptic revisions. Bearing surfaces were categorized as OZ-CPE, CoCr-HXLPE, or CoCr-CPE. HXLPE inserts were stratified according to brand name. Confounding was addressed using propensity score weights. Marginal Cox-regression models adjusting for surgeon clustering were used.Results The proportion of females was 62%. Average age was 68 (SD 9.3) years, and median follow-up time was 2.8 (IQR 1.2–4.9) years. After adjustments, the risks of all-cause, aseptic, and septic revision with CoCr-HXLPE and OZ-CPE bearings were not statistically significantly higher than with traditional CoCr-CPE bearings. No specific brand of HXLPE insert was associated with a higher risk of all-cause, aseptic, or septic revision compared to CoCr-CPE.Interpretation At least in the short term, none of the alternative knee bearings evaluated (CoCr-HXLPE or OZ-CPE) had a greater risk of all-cause, aseptic, and septic revision than traditional CoCr-CPE bearings.