Predicting patient reported outcome in total knee arthroplasty using body mass index and limb measurements

Body mass index (BMI) has not been shown to correlate with Patient Reported Outcome Measures (PROMs) following total knee arthroplasty (TKA). We investigated the relationship between weight, BMI, limb morphology, and Oxford Knee Score (OKS). Furthermore, the utility of a novel radiological measurement, the Knee Mass Index (KMI), was investigated.

Data including weight, BMI, gender, preoperative and 12 month OKS were collected from an arthroplasty database that contained 268 patients who underwent TKA. Measurements of soft tissue and bone width were made from the preoperative radiograph and 'KMI' was calculated. Pearson correlation and multivariate regression analyses were used to assess the relationship between OKS and the above variables.

The novel measurement, KMI, was not a predictor of the OKS. The BMI was predictive of initial OKS (Odds Ratio (OR) -0.26 p < 0.001), 12 month OKS (OR -0.39 p < 0.001) and change in OKS (OR -0.39 p < 0.001). The initial OKS was predictive of 12 month OKS (OR 0.32 p < 0.001) and change in OKS (OR -0.68 p < 0.001).

The novel KMI metric was not useful in predicting function. Both the post-operative OKS and change in OKS are predicted by BMI and pre-operative OKS. This is one of the first studies to show a relationship between BMI and OKS.