Dose normalization is usually done if the amount of the drug substance is different between test and reference formulations that are not intended to be identical in their release characteristics. (e.g. comparison between immediate release b.i.d. with extended release o.d.)

Usually both raw data (i.e. without dose normalization) and data after dose ormalization are presented.

Then the reasons then depend on the scope of testing.

I attach two papers that may provide useful information (both are open access).

M. Sardina Thanks for you response. Apologies I forgot to mention the doses from the different sources of vit C are different, eg, 1, 4, 5g of oral, and IV administration of vit C are being compared. Once I dose normalise everything to 1g ie divide Cmax of 4,5g ect by 4 and 5 respectively, could I then do Cmax per mg of vitamin c? Providing me with dose normalised Cmax comparisons on a mg per mg basis?

Thanks for your time.

Tony Tiller I guess you can do it the way you described provided that pk parameters (the actual values) show dose linearity in the range of doses tested. You might wait for other answers to confirm this approach. As a clinical pahramcologist I do have some knowledge of pharmacokinetics but I am not a pharmacokinetists!!

Best

Hi,

If you have different doses tested for a BA of course you have to dose normalize your PK parameters, i.e. devide on the dose given. This is applicable for dose related parameters, such as AUC, Cmax. Not applicable for time related parameters (tmax, t1/2) and for PK parameters calculated using dose already (CL, Vz).

An other advise, use geometric (or LS) means to compare and not average. Consequently, use CI instead of SD (or CV%) which is related to LS mean.