Response to ‘Sex‐by‐formulation interaction in bioequivalence studies: the importance of formulations and experimental conditions’ by Ibarra et al.

We are glad to read in the previous commentary that Ibarra et al agree that studies suggesting the existence of sex-by-formulation interaction should not be considered as demonstration of such an interaction if the study was not designed for that purpose and the sample size was not designed for that investigation.1 As regulators and scientists, we investigated the existence of a sex-by-formulation interaction in efavirenz tablets,2 as suggested by Ibarra et al,3 because we also consider this topic of interest. If the interaction really exists, the present bioequivalence requirements in regulatory guidelines would need to be revised worldwide. However, only a few papers have addressed this topic, and these papers are observational in nature and subject to bias.2 Consequently, it would be necessary that Ibarra et al conduct a confirmatory study with the formulation that suggested the existence of a sex-by-formulation interaction to obtain evidence from a properly designed experiment. On our side, we will continue exploring critically this possibility on the available data in order to improve the existing evidence, although observational. As argued by Ibarra et al, males are more likely to be in fasted state because digestion in males is more rapid than in females. Therefore, this suggests the different discriminatory power is not related to sex but to fasting/fed state. Consequently, even assuming that a sex-by-formulation interaction may exist, there is no need to demonstrate bioequivalence separately in males and females if bioequivalence is shown separately in fasted and fed state, as it is required by the US-FDA presently, since the underlying cause is the fasting/fed state of the study participants. Given there would be a wide range in “degree of digestion” in both the female and male populations at a time such as 2 hours after a meal, the continued study at the two ends of the range, under completely fasted conditions and immediately following a meal, seems like the best way to address the variability that would be observed in between. The influence of the extreme values was pointed out to illustrate that mean values of groups with less than 12 subjects are not reliable. We appreciate the notification of an error in subject #10 value, which is inconsequential in our view. After the exclusion of the extreme values, the means of the sex groups continue to be unreliable since the sample size is too small. Ibarra et al claim that a sex-by-formulation interaction arises only if the evaluated formulations are truly different and acknowledges that the suggested sex-by-formulation in our pilot study could have been an artefact because it was bioequivalent to the reference in the pivotal study. However, the pilot study reported in our paper included two test formulations. The formulation that was not tested in the pivotal study, because it was the most different formulation with respect to the reference, did not exhibit differences in the point estimates between males and females.2 As the sex-by-formulation interaction is dependent not only on the assayed formulations but also on the experimental design, as claimed by Ibarra et al, differences between sex groups were not observed because both males and females were in fasted state and, therefore, both sex groups were equally discriminative in our pilot study. Consequently, the sex-by-formulation interaction detected by Ibarra et al3 was created artificially by the tested method of administration, ie, 2 hours after dinner, which was insufficient to ensure that females (at least) were in a completely fasted state. It is important to highlight that it is true that in order to improve the tolerability of nervous system adverse reactions, bedtime dosing of efavirenz is recommended.4 However, it is recommended that efavirenz be taken on an empty stomach4 and this is not the case for this design in females according to Ibarra et al. The increased efavirenz concentrations observed following administration of efavirenz with food may lead to an increase in frequency of adverse reactions.4 Then, females should not receive the drug 2 hours after meal intake, which is also something to be confirmed because it is not considered during the clinical development of innovator products. In contrast to Ibarra et al, we consider that bioequivalence studies should be conducted in the most discriminative conditions and not mimicking the clinical setting because the diverse conditions of patient-use cannot be anticipated and the investigation of the worst-case scenario is better to ensure bioequivalence in all potential patients and scenarios. We commented on the results obtained by Ibarra et al3 to conclude that males and/or fasting state would seem to be more discriminative, if any, based on their results. Generalisation to include other results or the contrary results (eg, females being more discriminative because the fed state is more discriminative than the fasting state in other drug products) seems to be speculative, although hypothetically possible, because they are not based on the evidence of Ibarra et al. In contrast to Ibarra et al, we would propose the demonstration of bioequivalence in males and females separately, irrespective of the cost, which would simply be duplicated, if it were shown that the sex-by-formulation may really exist, because the absence of bioequivalence in one sex group is not ethically acceptable and a combined analysis is not able to ensure bioequivalence in both sex groups. Finally, the advancement of science is based on the demonstration that the null hypothesis (ie, bioequivalence conclusion is not different between sexes) is incorrect. Therefore, the alternative hypothesis of the existence of a sex-by-formulation interaction needs to be demonstrated formally to propose a change in the existing bioequivalence guidelines. We remain interested in investigating this possibility to avoid any risk to public health, and we appreciate and encourage scientific debate. There are no competing interests to declare.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.