## Fast & fed in one study (alternative) [Design Issues]

Hy everyone,

» » By “much more” you mean that the 90% CI is entirely outside the 80-125% limits, right?

» Not necessarily. Yesterday when I was asking the question I was thinking about 90% CI partially out, e.g., T fed/T fasting is 60%--85%, and R fed/R fasting is 115%--130%. So food increase the reference but decrease the test absorption. Or vice versa.

»

» But what if 90% CIs are within 80%--125%? e.g., T fed/T fasting 80.1%--90% (which is statistically significant different but is BE) and R fed/R fast about 115%--124% (similarly, there's statistically significant difference but is BE), or vice versa. Clearly, food effect is huge and there's something wrong here...

»

» Of course, this is just pure speculation. I'm not even sure if the scenario is remotely possible given that T fed/R fed and T fast/R fast are assumed within 80%--125%...

I think this scenario would be rare to happen. Unless the matrices of the generic and RLD formulations are totally different, impacting on excipient solubilitY under fed condition and, consequently, different drug release rate would be obtained (e.g., hydrophilic vs lipophilic matrix).

In any case, if the health authority will evaluate the two studies separately (if the two studies were perfomed as two independently studies:Fasting 2x2x2 and Fed 2x2x2 - scenario considered by ANVISA), I do not believe there is a problem with that if both studies were BE.

» » By “much more” you mean that the 90% CI is entirely outside the 80-125% limits, right?

» Not necessarily. Yesterday when I was asking the question I was thinking about 90% CI partially out, e.g., T fed/T fasting is 60%--85%, and R fed/R fasting is 115%--130%. So food increase the reference but decrease the test absorption. Or vice versa.

»

» But what if 90% CIs are within 80%--125%? e.g., T fed/T fasting 80.1%--90% (which is statistically significant different but is BE) and R fed/R fast about 115%--124% (similarly, there's statistically significant difference but is BE), or vice versa. Clearly, food effect is huge and there's something wrong here...

»

» Of course, this is just pure speculation. I'm not even sure if the scenario is remotely possible given that T fed/R fed and T fast/R fast are assumed within 80%--125%...

I think this scenario would be rare to happen. Unless the matrices of the generic and RLD formulations are totally different, impacting on excipient solubilitY under fed condition and, consequently, different drug release rate would be obtained (e.g., hydrophilic vs lipophilic matrix).

In any case, if the health authority will evaluate the two studies separately (if the two studies were perfomed as two independently studies:Fasting 2x2x2 and Fed 2x2x2 - scenario considered by ANVISA), I do not believe there is a problem with that if both studies were BE.

—

Marcelo Davanço

Marcelo Davanço

### Complete thread:

- Fast & fed studies Amira Gouda 2020-03-01 20:28 [Design Issues]
- Fast & fed studies alghazam 2020-03-03 10:24
- Fast & fed studies Ohlbe 2020-03-03 11:58
- Fast & fed in one study (alternative) Helmut 2020-03-04 14:15
- Fast & fed in one study (alternative) Shuanghe 2020-03-04 16:13
- Fast & fed in one study (alternative) jag009 2020-03-04 19:25
- Fast & fed in one study (alternative) Helmut 2020-03-05 11:50
- Fast & fed in one study (alternative) Shuanghe 2020-03-05 16:02
- Fast & fed in one study (alternative)Marcelo Davanco 2020-03-06 13:35

- Fast & fed in one study (alternative) Shuanghe 2020-03-05 16:02

- Fast & fed in one study (alternative) Shuanghe 2020-03-04 16:13

- Fast & fed in one study (alternative) Helmut 2020-03-04 14:15