Bioequivalence and Bioavailability Forum

Dear Ravi!

❝ Is there any regularity guidline about How to access the dose proportionality in a parallel and crossover biostudy?

As far as I know: no.

❝ Treatments [...] were (1). Test1 (one 100mg tablet of drug X), (2). Test2 (Two 100mg tablets of drug X), (3). Test3 (No tablet or drug X was given). Is it possible to access dose proportionality in this type of study.

If your drug is not an endogenous compound, I would expect to see no response at the zero dose level.
Most people would use a power model
  E(Y|x) = a · x b
or the log-transformed version
  log(E(Y|x)) = log(a) + b · log(x)
were Y is the untransformed PK-response (AUC, C_max) and x the dose. Regression is performed with weights set to 1/x. Bad luck if your compound is an exogenous compound (E(0) = 0) - weighting will will not work at x = 0. So you're in the trap of only two dose levels. You cannot distinguish between a linear model and a power model (or any other with two parameters). Furthermore you have no degrees of freedom left, in other words for any assessment of dose proportionality you will need at least three dose levels.
For details see one of my lectures.