Predicting implant size in total knee arthroplasty using demographic variables

Area of Science:

• Orthopedic surgery
• Biostatistics
• Medical device engineering

Background:

• Total knee arthroplasty (TKA) implant sizing is crucial for optimal patient outcomes.
• Accurate prediction of implant size can streamline surgical planning and inventory management.
• Current methods for TKA implant sizing rely on surgeon experience and intraoperative measurements.

Purpose of the Study:

• To investigate the relationship between patient height, weight, and sex with total knee arthroplasty (TKA) implant size.
• To develop and validate predictive models for TKA implant size selection.
• To assess the accuracy of multivariate linear regression and Bayesian models in predicting implant size.

Main Methods:

• Retrospective review of 8,100 primary TKAs from an institutional registry (2005-2016).
• Collected patient demographics: age, sex, height, weight, and BMI.
• Developed and validated multivariate linear regression and Bayesian models using training and testing cohorts.

Main Results:

• Height showed a strong correlation with implant size (femoral AP ρ = 0.73, tibial ML ρ = 0.77).
• Weight demonstrated a moderate correlation with implant size (femoral AP ρ = 0.46, tibial ML ρ = 0.48).
• Linear regression models explained significant variance in implant size (femoral R² = 0.607, tibial R² = 0.695).
• Bayesian model accurately predicted implant sizes (94.4% femur, 96.6% tibia) within a 5% inaccuracy tolerance.

Conclusions:

• Patient height, weight, and sex are significant predictors of TKA implant size.
• Developed linear regression and Bayesian models accurately predict required implant sizes.
• These predictive models offer potential for improved operating room efficiency and implant supply chain management.