NONMEM Users Network Archive

Hosted by Cognigen

RE: Extrapolation to achieve actual half-life

From: Kevin Feng <Kevin.Feng>
Date: Fri, 16 Mar 2018 16:02:21 +0000

Dear Anuja,

Just to clarify a few things:
"parallel study for BE- failed on the lower side of Cl limits"

- Solid oral administration

- Test formulation (T) dissolution profile may be slower than the re=
ference formulation (R)

- The PK observations is good enough to produce PK parameters such a=
s Cmax, Tmax, t_1/2, AUClast, AUCinf from NCA

- You already have a few in vitro dissolution profiles for both T an=
d R from different in vitro test methods e.g. USP apparatus, rotation speed=
, pH, volumes, media type etc.
"achieved final model and done with VPC,LLP "

- You use both T and R to build and validate the models.

- Or you build models from T and R separately

- It is a very good sign that the final model achieved with VPC etc.

With the above clarification, I would suggest to do steps below:
Model built and in vitro test method identification:

1. Find the in vitro dissolution models e.g. hill or Weibull

2. Add the in vitro model to your build pop-PK model - e.g. ref: Buchw=
ald, 2003 JPP

3. Validate the models with both R and T plasma/blood concentration ob=
servation data

4. Find the best validated in vitro test methods for simulation - e.g.=
 comparing OFV or AIC etc.

1. Used the validate models above. Note: if you use models built separ=
ately from T and R, it is very hard or impossible to do a 2X2 crossover sim=
ulation because you need to create same individuals every time for the cros=
sover design

2. Do the simulation with 2X2 crossover and make sure the sequence and=
 period correct and well design. (TR & RT)

3. Produce the PK parameters such as Cmax, Tmax, AUC

4. Do the BE test and find CI limits using the testing formulation fro=
m the simulated data
Test formulation validation from test bounds

1. Make sure you have enough washout at this step for all individuals =
in the simulation e.g. far more than 4~5 half life

2. If you T is successful in the BE test from the above 2X2 crossover =
simulation, you may want to see how your T is lay within the test bounds

3. Manually create a lower bound and upper bound in vitro profile

4. Build this in vitro models from these test bounds

5. Add this in vitro models for 2X2 crossover simulation and get the B=
E test result

6. Make sure your T is lay within the test bounds in vitro.

7. If you T is outside the test bounds in vitro, you may need to think=
 about to have a new formulation strategy. At this step, you need to make s=
ure you use the correct in vitro test methods with the above model building=
 step e.g. with correct USP apparatus, rotation speed, pH, volumes, media t=
ype etc.
Sufficient wash-out period design

1. Use the above validated models for 2X2 crossover simulation

2. Manually create a few scenarios with different wish out period

3. Do the BE test from all the above scenarios - better to do a plot t=
ogether with (80,125) lines

4. Do the ANOVA or linear mixed effect analysis to make sure no carry =
over effect e.g. formulation: sequence etc.

5. Find the best scenario and shortest wish out period with the succes=
sful BE test without carry over effect.

We have a tool available to automate all the above processes. If you happen=
 to attend PAGE meeting this year, you can find us from Certara booth and w=
e can do you a demo.
               Automatic framework for bioequivalence studies from In Vitro=
 test to In Vivo study design

Best wishes,
Kevin Feng

From: owner-nmusers
Sent: Wednesday, March 14, 2018 1:17 AM
To: nmusers
Subject: [NMusers] Extrapolation to achieve actual half-life

Hello Everyone,
We have done one parallel study to prove Bioequivalence, which failed on th=
e lower side of Cl limits. We want to do crossover study, but in previous s=
tudy we failed to capture actual half life. The API is having a half life a=
round 4-5 days and we did a parallel study for 2 days.

By M&S, we want to extrapolate to get actual half-life. And we want to calc=
ulate the sufficient wash-out period so that the carry-over effect will be =
<5% of Cmax in period 2 of crossover study.

We have achieved final model and done with VPC,LLP, please guide me for sim=

Thanks & Regards
''Legally privileged confidential information and subject to "Disclaimer". =
<> NOTICE: The infor=
mation contained in this electronic mail message is intended only for the p=
ersonal and confidential use of the designated recipient(s) named abov=
e. This message may be an attorney-client communication, may be protected=
  by the work product doctrine, and may be subject to a protective orde=
r. As such, this message is privileged and confidential. If the reader=
 of this message is not the intended recipient or an agent responsible for =
delivering it to the intended recipient, you are hereby notified that =
you have received this message in error and that any review, dissemina=
tion, distribution, or copying of this message is strictly prohibited. If y=
ou have received this communication in error, please notify us immedia=
tely by telephone and e-mail and destroy any and all copies of this me=
ssage in your possession (whether hard copies or electronically stored copi=
es). Thank you. buSp9xeMeKEbrUze

Received on Fri Mar 16 2018 - 12:02:21 EDT

The NONMEM Users Network is maintained by ICON plc. Requests to subscribe to the network should be sent to:

Once subscribed, you may contribute to the discussion by emailing: