Exponential heating in drug stability experiment and statistical evaluation of nonisothermal and isothermal prediction.

A new nonisothermal heating model (exponential heating model) for drug stability experiments, based on a theoretical study of simulated nonisothermal data, is presented. In the model, the heating rate dT/dt is increased by 2-4 times at every increase of 10 degrees C in temperature: dT/dt = a(T-T0)/10.(dT/dt)0 where a is the times by which the heating rate is increased at every increase of 10 degrees C. A computation method with optimization and Simpson integration for the experiment was also introduced. The estimates for the shelf-life and activation energy obtained by the exponential and other nonisothermal heating models were statistically evaluated and were compared with those obtained by the isothermal method under various conditions. The results indicated that under the same experimental conditions, the estimates obtained by the exponential heating model were significantly more accurate and precise than those obtained by the linear, reciprocal, and logarithmic heating models. The accuracy and precision of the estimates were independent of the shelf-life of drugs and experimental period. The estimates obtained by the isothermal method were somewhat more accurate and precise than those obtained by the exponential heating model. However, the experimental period needed by the isothermal method was about five times longer than that needed by the exponential heating model. The results also showed that in each of the heating models, the estimates are more accurate and precise by increasing the extent of drug degradation, changing temperature range or sampling frequency, or by having the mean temperature closer to room temperature. To demonstrate its applicability, the exponential heating model was used to study the stability of vitamin C tablets and predict their shelf-life.