Abstract
Purpose

Periprosthetic joint infection (PJI) is a major cause of revision surgery after total knee arthroplasty (TKA) and unicondylar knee arthroplasty (UKA). Patient- and hospital-related risk factors need to be assessed to prevent PJI. This study identifies influential factors and differences in infection rates between different implant types.
Methods

Data were obtained from the German Arthroplasty Registry. Septic revisions were calculated using Kaplan–Meier estimates with septic revision surgery as the primary endpoint. Patients with constrained and unconstrained TKA or UKA were analysed using Holm's multiple log-rank test and Cox's proportional hazards ratio. The 300,998 cases of knee arthroplasty analysed included 254,144 (84.4%) unconstrained TKA, 9993 (3.3%) constrained TKA and 36,861 (12.3%) UKA with a maximum follow-up of 7 years.
Results

At 1 year, the PJI rate was 0.5% for UKA and 2.8% for TKA, whereas at 7 years, the PJI rate was 4.5% for UKA and 0.9% for TKA (p < 0.0001). The PJI rate significantly increased for constrained TKA compared to unconstrained TKA (p < 0.0001). The PJI rate was 2.0% for constrained TKA and 0.8% for unconstrained TKA at 1 year and 3.1% and 1.4% at 7 years. Implantation of a constrained TKA (hazard ratio [HR] = 2.55), male sex (HR = 1.84), increased Elixhauser score (HR = 1.18–1.56) and implant volume of less than 25 UKA per year (HR = 2.15) were identified as risk factors for revision surgery; an Elixhauser score of 0 (HR = 0.80) was found to be a preventive factor.
Conclusions

Reduced implant volume and constrained knee arthroplasty are associated with a higher risk of PJI. Comorbidities (elevated Elixhauser score), male sex and low UKA implant volume have been identified as risk factors for PJI. Patients who meet these criteria require specific measures to prevent infection. Further research is required on the potential impact of prevention and risk factor modification.
Level of Evidence

Level III.