ABSTRACT
Long-term ingestion of fluoride through food and drinking water sources cause dental, skeletal, and non-skeletal manifestations. Severe grades of skeletal fluorosis are indicated to be an irreversible process in the human body. However, few medical studies especially in children are available indicating the feasibility of its treatment and reversal, even though no mathematical analysis or modeling has been attempted by these researchers. Three published Indian double-blind clinical studies on the reversal of skeletal fluorosis in children has been analyzed to investigate the effects of increased fluoride exposure, and develop a predictive mathematical model for predicting the doses of calcium which is crucial for reversing the effects of fluoride ingestion. Bivariate analysis was used for quantitatively representing the steps involved in the pathophysiology of fluoride intake in human body. The probability of skeletal fluorosis occurrence was classified on the basis of physical parameters namely: age, weight, and fluoride intake. Multilayer perceptron (MLP) technique which adopted the neural network approach was used to obtain the normalized importance of various biochemical parameters. The 196resulting significant parameter, serum GAG, along with calcium retention equation was used to develop a model to predict the doses of calcium which can reverse the effects of fluoride named as reduced retention model (RRM). The calcium doses predicted by RRM at different drinking fluoride levels were in coherence with the doses recommended by clinicians and therefore predicted doses could be used as guidelines for clinical recommendations. The results of MLP analysis also support the possibility of the use of serum GAG as a parameter in diagnosing fluorosis at an early stage in developing children, which may prove useful in improving the efficacy of treatment.
