Prediction of long-term outcome of tibial osteotomy in medial gonarthrosis.

To predict the long-term outcome after tibial osteotomy for medial gonarthrosis, multivariate statistical techniques were used and prognostic equations were defined. A comparison between different evaluation systems was also performed. Clinical and radiographic data from 81 patients followed up for more than 7 years were used. Twenty-eight preoperative variables were analysed in relation to the outcome at 7 years with the multivariate procedures discriminant analysis and cluster analysis. Among the variables, knee flexion and walking capacity were found to be the most dominant factors. Excluding the undercorrected osteotomies, preoperative walking capacity was now the dominating variable for prediction. Seven years postoperatively only small differences were found between the evaluation criteria of pain relief, London Hospital score, and HSS score. The discriminant functions for the highest ranked evaluation criteria are given in the text, as are dendrograms showing the patterns of interdependence between the evaluation criteria. Thus it was possible to transform clinical entities into factors suitable for determination and calculation of postoperative evaluation. The most commonly used evaluation systems had very similar discriminant functions showing good agreement.