Statistical approaches to indirectly compare bioequivalence between generics: a comparison of methodologies employing artemether/lumefantrine 20/120 mg tablets as prequalified by WHO.

PURPOSE

The objective of this study was to compare different methods of adjusted indirect comparisons that can be used to investigate the relative bioavailability of different generic products. To achieve this goal, generic artemether/lumefantrine 20/120 mg tablets that have been prequalified by the World Health Organization (WHO) were selected as model products for study.

METHODS

Data from three bioequivalence studies conducted independently that compared three generics with the same reference product were used to indirectly determine the relative bioavailability between the generics themselves.

RESULTS

The different methods of indirect comparison examined in this study provide consistent results. Methods based on the assumption of a large sample size give slightly narrower 90 % confidence intervals. Therefore, the use of methods based on the t test is recommended. Given the precision of the area under the time-concentration curve (AUC) data, it is possible to conclude that the extent of exposure of artemether and lumefantrine is bioequivalent between the different generics studied. However, given the precision of the drug peak concentration (C(max)) data, it is not possible to demonstrate equivalence within the conventional acceptance range for all comparisons; it is possible to conclude bioequivalence within the widened acceptance range 75-133 %.

CONCLUSIONS

From a clinical viewpoint, not only are these prequalified generics bioequivalent and interchangeable with the reference product (Coartem, Novartis), but also the existing indirect evidence makes it possible to conclude that these WHO prequalified products are bioequivalent between themselves with respect to the AUC. The lack of the necessary precision to demonstrate bioequivalence between generics with respect to the C(max) within the conventional acceptance range does not preclude considering them as interchangeable, if necessary, since C(max) is considered to be of less clinical relevance for the relevant therapy.