ICH drug stability research evaluation

 

The purpose of the ICH drug stability study is to establish, based on the test of at least three batches of drug substance stability and evaluation information (including physical, chemical, biological and microbiological tests with appropriate results), the next period of retesting is applicable to all batches.

The data may show little degradation and little variability, so it is obvious from looking at the data that the required retest period will be approved. In this case, there is usually no need for a formal statistical analysis; it is sufficient to provide justification for the omission.

One way to analyze data for quantitative attributes expected to change over time is to determine the time when the 95% one-sided confidence limit of the average curve intersects the acceptance criterion. If the analysis shows that the variability from batch to batch is small, then it is advantageous to incorporate the data into a single overall estimate. This can be achieved by first applying an appropriate statistical test (for example, rejecting the p-value of the significance level greater than 0.25) on the slope of the regression line and the zero time intercept of a single batch. If it is not suitable to combine data from several batches, the overall retest cycle should be based on the minimum time a batch is expected to remain within the acceptance criteria.

The nature of any degenerate relationship will determine whether the data should be converted to linear regression analysis. Generally, this relationship can be expressed by linear, quadratic, or cubic functions on arithmetic or logarithmic scales. Statistical methods should be used to test the goodness of fit of the data from all batches and combined batches (where appropriate) to the assumed degradation line or curve.

If necessary, limited extrapolation of real-time data of long-term storage conditions beyond the observation range can be performed at the time of approval to extend the retest period. This proof should be based on the known degradation mechanism, test results under accelerated conditions, goodness of fit of any mathematical model, batch size, existence of supporting stability data, etc. However, this extrapolation assumes that the same degradation relationship will continue to apply outside the observed data.

Any evaluation should include not only determination, but also the level of degradation products and other appropriate attributes.