A model of a controlled drug release from tablets using neural network

A model of a controlled drug release from tablets using neural network
bahareh etaati,^1,* Hadi Ezatpanahniaki,² fereshteh fouladi,³ Fatemeh Ehsani Besheli,⁴
1. Department of computer Engineering, Amirkabir University of Technology
2. Department of Cumputer Engineering, Amirkabir University of Technology
3. Department of Chemical Engineering, Amirkabir University of Technology
4. Department of Cumputer Engineering, Amirkabir University of Technology
Introduction: Capsule capsules are an oral controlled drug delivery system. These pills alter the secretion of the drug by using insoluble bands that are used peripherally on the pills. A new two-dimensional model has been developed to predict drug release from these systems. Erosion occurs in several stages, but in most cases the destruction of the control stage appears to be the first-order reaction. In this study, the tablets are modeled and compared with the number, width, and position of the different bands. The results show an almost constant release rate for the 2-band tablet, which is close to zero order release. This model can be expressed using a neural network.
Methods: There have been several attempts to model erosion. We can classify these models into three groups: phenomenological (the models that are based on governing equations such as reaction, diffusion or dissolution), probabilistic (which model the changes in a system without an equation that describe the phenomenon and mostly Monte Carlo or Cellular automata methods have been used), and empirical (often these models are the first step toward another ones). [3] Phenomenological A dissolution model for different geometries, was developed by (Cooney, 1972) where the dissolution of a nonporous solid in a fluid involves 2 main steps: [5] Detachment of an atom, molecule, or ion from the solid surface, followed by diffusion of this species through an adjacent stagnant layer out into the bulk of the fluid. The solution was simplified for solids having a very slow dissolution rate that surface detachment tends to control the process. So the rate of released drug is: dw/dt = k. s. ∆c
Results: A finite difference method has been developed for solving equation 3 with moving boundary (equation 4). First mesh dependency has been checked by using different mesh sizes. Figure 4 shows Mt which is total amount of drug released until time t. As it is assumed that diffusion of drug in polymer is so slow and negligible compared to erosion rate, so Mt is calculated from equation 10. Results show Mt does not change with mesh size finer than 80×400 (figure 4). So in all simulations this mesh size has been used. Mt = Cd . Vte Cd is concentration of drug in solid polymer which is assumed to be homogenous. Vte is total volume of eroded parts of polymer until time t. Vte = π ∑N T=0 (1 – step ( cm - ʃt k Cw .dt)) .ri dri dzi step = 1 u > 0 0 u < 0
Conclusion: Finally to achieve a better zero order release, 2 bands tablet was modified and their width was A two dimensional model for drug release in Ring Cap tablets is developed. In this simulation it is assumed that the only mechanism of drug release is erosion. The parameters that can affect release rate are number, width, and placement of bands. According to the aim of controlling release rate different situations can be chosen. Tablet with two bands can create a constant release rate. It is so beneficial due to the constant amount of drug that is released over time. One band tablet has advantage of releasing drug in a longer time (long acting), so patients would use less dosage.Further research can be done to investigate the effect of different tablet geometries, and extra trials on changing number and width of bands to achieve better zero order control release rate. Vte = π ∑N T=0 (1 – step ( cm - ʃt k Cw .dt)) .ri dri dzi step = 1 u > 0 0 u < 0
Keywords: Erosion; Degradation; Ring Loop; Hydrolysis; Mathematical Model; Neural Network Model