PREDICTING THE OUTCOME OF TOTAL KNEE ARTHROPLASTY FROM PREOPERATIVE STATUS

Abstract

This paper aims to determine if preoperative characteristics have a significant impact on functional outcome as measured by the WOMAC at 2-years following total knee replacement (TKR) surgery. Patients were recruited as part of a prospective study of outcomes of primary TKR for osteoarthritis in centres in the US (4 centres), UK (6 centres) and Australia (2 centres). Research assistants recruited eligible patients and collected clinical history and physical examination data preoperatively, 3, 12 and 24-months post surgery. The WOMAC, SF-36, patients satisfaction and demographic data were obtained by self-administered questionnaires. All scores were transformed to 0–100 scale (100 best).

We recruited 860 eligible patients and have complete 12-month WOMAC data on 736 patients (86%) and 2-year data on 701 patients (78%). Mean age was 70 years (SD 10), 59% were female, 50% were from the UK, 30% from the US and 20% from Australia. Mean preoperative clinical measures were: knee flexion 107° (SD18), SF36 Mental Health 72 (SD19), body mass index 29 (SD 6) and WOMAC Function 45 (SD 19). 46% of patients reported more than 2 comorbid conditions.

There was no significant difference between mean WOMAC Function scores at 12-months (73, SD 21) and 2-years (74, SD21). In a linear regression model (model R- square= 25), the preoperative predictors of worse WOMAC Function at 2-years, in order of decreasing importance, were: low WOMAC Function (p< 0.0001), higher number of comorbid conditions (p=0.0002), UK patients (p=0.0002), low SF36 Mental Health (p=0.01) and restricted preoperative knee flexion (p=0.02). Patients who come to surgery with poor function, restricted knee flexion, low mental health and other comorbid conditions are more likely to have worse functional outcomes 2-years following surgery. After adjusting for these predictors, the UK patients had significant lower WOMAC Function scores than the US and Australia.