Validation of internal normalization for impurity assays

The simplicity of the internal normalization made it a very attractive meth od. Yet, because of its restrictive applicability requirements, internal no rmalization is not widely implemented in HPLC quantitative analysis. Basica lly, applicability requirements are that all the solutes must not only be e luted and detected but must also present similar behavior toward the detect ion system. Ideally, response factors should be identical for all the solut es or, in practice, of the same order of magnitude. The methodology develop ed to validate, in a rigorous way, internal normalization was based on the use of a statistical tool called analysis of covariance (ANACOVA). ANACOVA is more or less similar to ANOVA but can manage a continuous variable, like for example, concentration. So, it is possible to use it to compare calibr ation curves of all the different solutes present in a sample, for example, the main product and its impurity. After having checked that for the main product the response factor was the same around the target concentration of the HPLC method, and at low concentration, it was then possible to make co mparison with impurity behavior, and to determine whether the use of the re sponse factor was necessary or not. Eventually, ANACOVA enabled the validat ion of internal normalization by assessing that all the solutes presented r equired behavior. This methodology was successfully applied to an actual ex ample of liquid chromatography quantitative analysis, taken from the pharma ceutical industry. In this case, internal normalization for impurity assays of an anticytomegalovirus drug substance was validated after response fact or correction.