A linear validation method of analytical procedures based on the double logarithm function linear fitting

Abstract

BackgroundThe ICH Q2(R1) guideline defines linearity as its ability to obtain test results which are directly proportional to the concentration of the analyte. However, the coefficient of determination typically used is limited to evaluating the response function rather than the linearity of results, the current guidelines fail to provide a method for assessing the linearity of results. ResultsThe paper presents a data analysis method to validate the linearity of results. The method demonstrates the degree of data proportionality by applying double logarithm function linear fitting and solves the problem of setting the acceptance criterion by investigating the relationship between the slope, working range ratio, and maximum error ratio. In principle, this method has advantages over the coefficient of variation and recovery rate recommended by the ICH M10 guideline. Moreover, the relative error data show that the double logarithm function is more effective in overcoming heteroscedasticity than straight-line fitting. SignificanceThis method is more consistent with the linear definition outlined in ICH Q2 guidelines and has the potential to further clarify the concept and validation method of linearity in the future.