Bioequivalence and Bioavailability Forum

Dear MGR!

❝ If suppose male and female volunteers are taken for the study. In that case, can we take gender effect in the statistical analysis (ANOVA) or not? If not please explain. and even we have to take equal number?

The FDA specifically ask for females/males (general population), but don’t expect the study to be powered for each subgroup:

This guidance recommends that in vivo BE studies be conducted in individuals representative of the general population, taking into account age, sex, and race. We recommend that if the drug product is intended for use in both sexes, the sponsor attempt to include similar pro­por­tions of males and females in the study. […] We recommend that the total number of subjects in the study provide adequate power for BE demonstration, but it is not expected that there will be sufficient power to draw conclusions for each subgroup. Statistical analysis of subgroups is not recommended.

As you might know ANVISA’s website is a mess. The most recent document in English I have is Resolution Nº 896 (May 29, 2003).

[…] the studies could be conduct in subjects between 18 and 50 years of age and capable of expressing their free and informed consent, male, female or both, noting that in the last case, the number of men and women must be the same and equal distributed between the sequences.

The consolidated norms (July 2010) are only available in Portuguese; in page 50 you may find:

k) de acordo com o medicamento, os estudos poderão ser conduzidos em voluntários com idade superior a 18 anos e capazes de fornecer seu consentimento livre e esclarecido, do sexo masculino, feminino ou ambos, sendo que neste último caso, recomenda-se que o número de homens e de mulheres seja distribuído igualmente entre as seqüências. Se o medicamento for indicado para pacientes com aracterísticas específicas de idade e sexo, o estudo deverá ser integralmente realizado em voluntários com essas características. No caso de estudos para contraceptivos, o estudo deve ser realizado com mulheres em idade fértil;

As far as I understand () the rest of the document a statistical analysis is not required, though I have heard that ANVISA sometimes (!) asked for separate BE analyses of males/females. Don’t know why because power will be inferior. Maybe they were only interested in the T/R-ratio. No idea.

In my studies I try to recruit for a 50:50 ♀♂ ratio, but got never any complaints if the final ratio (especially due to drop outs in small studies) was just 70:30. I did/do not include sex as an effect in ANOVA.

Hi,

since the vast majority of subjects do not change gender between period 1 and period 2, I think it is quite reasonable to consider gender a between-subject factor. Thus, for the common crossover designs I would not expect it to mean a lot.
Strictly, I guess we can speak of gender-in-sequence depending on the coding. The Silly-o-meter might need an update

EM.