Bioequivalence and Bioavailability Forum

Dear Srinivasu,

have you read my original reply (given below)?

❝ ❝ It’s only possible to predict BE if you have established an in vivo-in vitro-correlation (IVIVC).

❝ Let me know the chances of bioequivalence.

No idea; I don’t have a crystall ball and I’m not a friend of reading tea leaves. It’s like you want to know something about the dark side of the moon and are equipped with a telescope only. But you need a space probe: Perform a (pilot) BE study.

Concerning your first question on other methods to compare dissolution profiles: You have an asymetric sigmoidal function. Based on your data you may fit e.g. a 4-parameter logistic function Y=(-dissinf/(1+(x/inflect)^slope)^asym)+dissinf, but what will you gain? You have only 5 degrees of freedom. Parameters (95% CI) are:
parameter     test                                 reference
dissinf   101.158771 (   24.182645, 178.134897)  94.688213 ( 83.270663, 106.105764)
inflect    40.030919 (-2386.12648, 2466.18832 )   6.046716 (-19.545979,  31.639411)
slope       1.509877 (    0.073172,   2.946581)   2.411028 (  0.740197,   4.081859)
asym       19.762344 (-1717.06569, 1756.59038 )   4.301083 (-31.747713,  40.349879)

Though the reference shows a ‘better’ fit, CIs of many parameters include zero and therefore are noninformative (aka nonsense).

Does this image show us anything new (what we haven’t seen by f₁=19.2% & f₂=43.4)? Obviously not.

If you want to start serious formulation development, you should start (!) with a small three-way cross-over study (test, reference, i.v. or an oral solution). Afterwards (!) try to develop a dissolution method which allows to establish an IVIVC (preferably Level A by deconvolution). Only then, you may peek at The Dark Side of the Moon…