Session I - Upper Extremity
Prediction of Instability of Fractures of the Distal Radius
Paul J. MacKenney, FRCS; Margaret M. McQueen, MD; Rob Elton, PhD, Edinburgh Orthopaedic Trauma Unit, Edinburgh, Scotland
Purpose: To produce a method of accurately predicting instability of distal radius fractures on the day of injury which is easily applicable in the clinical situation.
Methods: During a five-and-one-half year period, data were recorded prospectively on 4025 patients with distal radial fracture. The items of information included demographic data, mode of injury, the patient's level of function, and radiological data. The latter included: fracture classification (AO and Frykman); the degree of fracture comminution; measurement of dorsal angulation, radial shift and ulnar variance at presentation, post manipulation, one week, and six weeks after injury; carpal alignment at union. Using the position of the fracture at one and six weeks and the alignment of the carpus as outcome measures, statistical analysis of the data was performed to identify factors of independent significance in the prediction of early instability, late instability, and carpal malalignment respectively. Each factor was analysed using the Mann-Whitney or chi-squared test, and significant factors used in multiple logistic regression analysis. Predictive equations were then derived from the multiple logistic regression.
Results: Factors predictive of fracture instability were seen to vary depending on the position of the fracture at presentation (displaced with more than 10o of dorsal angulation or minimally displaced with less than 10o of dorsal angulation). There were also differences between factors predictive of early and late instability. The most highly significant prognostic factor was the age of the patient (p<0.001), and this applied to all sets of data analysed. Comminution was also important, though to a lesser degree. The dorsal angulation of the fracture at presentation was only of prognostic significance in fractures with minimal displacement at presentation, whereas the ulnar variance was significant in all fractures. Surprisingly, independence in the activities of daily living was of value in predicting late instability in displaced fractures. The sex of the patient, mode of injury, the radial shift at presentation, and Frykman classification was of no prognostic value following multiple logistic regression. In the prediction of carpal malalignment the patient's age and level of function, fracture comminution, AO and Frykman classifications, and the dorsal angulation of the fracture at presentation were all of significance.
From the data analysed it was possible to construct an equation to predict the percentage chance of instability after distal radius fracture. The following example is for the probability of early instability in a displaced fracture at presentation:
x = 0.03 x age + 0.38 (if comminution present) + 0.21 x ulnar variance - 3.12
Probability of instability(%) = [ex]/[1+ex]
An example can be given as follows. A 75-year-old lady sustains a distal radial fracture after slipping on ice. She is normally independent and active. Initial dorsal angle is 17 degrees and ulnar variance is 3 mm with dorsal comminution. This patient therefore has a fracture which is displaced at presentation. To calculate the probability of displacement at one week, the formula is used as above. The score is 0.14. Using the probability formula, this patient has an 82% chance of the fracture being unstable at one week.
Discussion and conclusions: Previously it has been demonstrated that the ulnar variance, dorsal angulation, and age of the patient were all factors related to the anatomical outcome following distal radial fracture. We have shown that the significance of these factors depends on the position of the fracture at presentation, and whether one is trying to predict early or late instability. We have shown that the age of the patient is the by far the most important factor in assessing the instability of a fracture. It is hoped that refinement of the predictive equations will produce a user-friendly and reliable method of assessing the distal radial fracture at presentation. The value of such a system cannot be underestimated: unnecessary manipulations can be avoided in patients with fractures that require operative fixation, and these patients will be able to have definitive treatment at presentation. In addition, this rationalisation of the treatment protocol is likely to be of economic benefit.