Background: Periprosthetic infection following total knee replacement remains alarmingly frequent, is devastating for patients, and is an increasingly substantial public health burden. While both the prosthetic design elements and type of polyethylene used in total knee replacement can independently affect survivorship of the prosthesis, their influence on the risk of revision for infection is unknown. Therefore, we examined the effects of total knee prosthetic design and type of polyethylene bearing on the long-term revision risk for nonacute infection following total knee replacement by comparing 4 different cohorts: patients who had minimally stabilized total knee prostheses with crosslinked polyethylene (XLPE) bearing surfaces, minimally stabilized total knee prostheses with noncrosslinked polyethylene (NXLPE), posterior stabilized total knee prostheses with XLPE, and posterior stabilized total knee prostheses with NXLPE.

Methods: National registry data on revision procedures for surgeon-reported infection following primary total knee replacement for osteoarthritis from September 1999 through December 2015 were obtained for 4 patient cohorts. The cohorts were defined by the total knee prosthetic design and polyethylene type used. Revisions at <=6 months were censored to reduce confounding bias. Hazard ratios (HRs) were adjusted for age, sex, and antibiotic cement usage.

Results: A total of 336,997 primary total knee prostheses were included, 1,651 (0.49%) of which underwent revision for periprosthetic infection. Compared with minimally stabilized total knee prostheses that had XLPE bearing surfaces, the revision risk for infection, adjusted for age, sex, and antibiotic cement usage, was 25% higher for minimally stabilized total knee prostheses with NXLPE bearing surfaces (HR = 1.25 [95% confidence interval (CI), 1.07 to 1.45]; p = 0.003), 89% higher for posterior stabilized total knee prostheses with XLPE (HR = 1.89 [95% CI, 1.52 to 2.35]; p < 0.001), and 102% higher for posterior stabilized total knee prostheses with NXLPE (HR = 2.02 [95% CI, 1.72 to 2.37]; p < 0.001). Posterior stabilized total knee prostheses with NXLPE had a 61% higher risk of infection compared with minimally stabilized total knee prostheses with NXLPE (HR = 1.61 [95% CI, 1.43 to 1.83]; p < 0.001). The revision risk for infection for posterior stabilized total knee prostheses with NXLPE was the same as that for posterior stabilized total knee prostheses with XLPE (HR = 1.08 [95% CI, 0.88 to 1.32]; p = 0.481).

Conclusions: Minimally stabilized total knee prostheses with NXLPE bearing surfaces and posterior stabilized total knee prostheses, irrespective of bearing type, had a greater long-term revision risk for periprosthetic infection when compared with the revision risk for minimally stabilized total knee prostheses with XLPE bearing surfaces.