Estimating the weight of dental amalgam restorations.

AIM Data on the weights of amalgam restorations are lacking. The aim of this study was to determine these weights and to develop criteria to facilitate their estimation. METHODS Four separate regression models with 4 covariates in various combinations were used to estimate the weight of amalgam restorations. Model I, based on 514 restorations from both natural and anatomical replica teeth, contained 3 covariates: the number of restored surfaces (covariate A), the type of tooth (covariate B) and whether the restoration had been removed from a natural tooth or an anatomical replica tooth (covariate C). Model II, based on 359 restorations from anatomical replicas, contained 2 covariates: A and B. Model III, based on 155 restorations from natural teeth, contained 3 covariates: covariates A and B and whether the natural teeth had been extracted in 2002 or at least 15 years previously (covariate D). In model IV, covariate D was removed from model III. RESULTS Model I explained 72% of the variation in the weight of restorations; the partial R2 for covariates A, B and C in model I was 0.5818, 0.797 and 0.0579, respectively (p < 0.001). In model III, the weights of the restorations did not depend on covariate D (p = 0.93). The least square mean weight of amalgam restorations with 1, 2, 3, and 4 or more surfaces restored (and 95% confidence interval) was 0.31 g (0.28-0.34 g), 0.49 g (0.45-0.53 g), 0.81 g (0.76-0.86 g) and 1.38 g (1.31-1.45 g), respectively. CONCLUSION The number of surfaces restored (covariate A) accounted for at least 80% of the variation in the weight of restorations in all models and therefore provides the best estimate for the weight of amalgam restorations.