Abstract
AIM OF THE STUDY: The aim of this case-control study was to develop a clinical decision rule to support assessment of the risk of long-bone non-union and plan for appropriate early intervention. METHODS: Two hundred patients (100 cases and 100 controls) were recruited. Risk factors identified to contribute to the development of non-union were recorded and analysed with a multivariable logistic regression model. Tabulation of the outcome (non-union/union) against each risk factor in turn (univariable analysis) was carried out. Odds ratios and confidence intervals were derived using Wald's method. A receiver-operator curve was calculated and the area under the curve was computed. Having established the eight most important risk factors, a non-union risk index was developed as the count of the risk factors present in each patient. RESULTS: The five risk factors for non-union with greater effect size were post-surgical fracture gap > 4 mm (odds ratio (OR) = 11.97 95% CI (4.27, 33.53)), infection superficial/deep (OR 10.16 (2.44, 42.36)), not optimum mechanical stability (OR 10.06 (3.75, 26.97)), displacement > 75% of shaft width (OR 6.81 (2.21, 20.95)), and site of fracture-tibia (OR 4.33 (1.32, 14.14)). The ROC curve for the non-union index was 0.924, sensitivity 91%, specificity 77%. CONCLUSIONS: The non-union index derived from counting risk factors predicts union for 0-4 risk factors and non-union for 5-8 risk factors. It can be readily applied and can guide clinicians about the risk of development of long-bone non-union. It can become a powerful aid for assessing fracture fixation outcome and to support early intervention.