Evaluation of non-isothermal methods in stability studies of human insulin pharmaceutical preparations

The purpose of this research was to study the thermal stability of a human insulin pharmaceutical preparation using non-isothermal conditions and comparison with classical isothermal experiments. The isothermal studies were performed in the temperature range 20–60 °C, whereas non-isothermal stability studies were performed using a linear increasing temperature program, heating rate 0.25 °C per hour and temperature interval 30–70 °C. Under isothermal conditions, an apparent first-order degradation process was observed at all temperatures. The linear Arrhenius plot suggested that the insulin degradation mechanism was the same within the studied temperature range, with quite large uncertainties due to the small number of degrees of freedom based only on the scatter in the plot, giving an estimated shelf-life at 25 °C of 199.1 days. In non-isothermal conditions, the integral approach was used to estimate the activation parameters. It provides results in good agreement with those of the traditional method, but with the advantage that the uncertainty in the final result directly reflects the goodness of fit of the experimental data, since it takes into account the scatter in the original data. The estimated shelf-life in non-isothermal conditions was quite close to the value derived from isothermal data, 191.4 days, although the 95% confidence interval estimated were slightly higher. This is due to the differences in the estimation method and the nature of the experimental errors. The bootstrap technique is also applied to estimating confidence limits for the Arrhenius parameters and shelf-life. This method is very useful when the underlying distribution function of the parameters is unknown. The results obtained indicate that the Arrhenius parameters follow a normal distribution, whereas the shelf-life follows a log-normal distribution. In any case, the results obtained show that there is no difference between the asymptotic and bootstrap confidence intervals.