mechanistic ivivc using the simcyp adam model -...

Mechanistic IVIVC Using the Simcyp ADAM

Model

Make SCIENCE out of IVIVC

© Copyright 2012 Certara, L.P. All rights reserved.

IVIVC and Its Components

0

50

100

150

-6 4 14 24

Pla

sm

a C

on

c

(ng

/mL

)

Time (h)

Formulation PK

0

20

40

60

80

100

0.0 4.0 8.0 12.0

%D

isso

lved

/Ab

so

rbe

d

Time (h)

In vivo Absorption/Dissolution

020406080

100

0.0 4.0 8.0 12.0

%D

isso

lved

Time (h)

In vitro Dissolution

y = 0.9941x R² = 0.9703

0

50

100

0 50 100

in v

ivo

%

Dis

so

lved

/Ab

so

rbed

in vitro % Dissolved

IVIVR

IVIVC: A predictive mathematical model describing the relationship between an in vitro property (usually drug dissolution or release) and a relevant in vivo response, e.g., amount of drug absorbed.

DECONVOLUTION

IVIV Relationships: The in vitro dissolution and in vivo input curves may be directly superimposable or may be made to be superimposable by the use of a scaling factor/function. Nonlinear correlations, while uncommon, may also be appropriate.

IVIVR

MODELS

MODELS

0

20

40

60

80

100

0.0 4.0 8.0 12.0Time (h)

IVIVR Predicted in vivo Abs

0

50

100

150

-6 4 14 24Time (h)

Predicted PK CONVOLUTION

VALIDATION


IVIVC Components: Deconvolution Methods

Deconvolution Methods

Model Dependent Methods Model Independent Methods

Wagner Nelson

Algebraic Equation Based Differential Equation Based

Loo Riegelman

DE Compartmental Model

Physiological (ADAM) Model

Direct Numerical Deconvolution

Numerical Deconvolution Through Convolution

Deconvolution in simplified terms is the method to estimate the rate of input in the system that produces the observed response (PK profile)

0

20

40

60

80

100

120

0 10 20

Pla

sm

a C

on

c (

ng

/mL

)

Time (h)

Formulation PK

0

20

40

60

80

100

0.0 4.0 8.0 12.0

%D

isso

lved

/Ab

so

rbed

Time (h)



Deconvolution: Mechanistic ADAM Model

in vivo dissolution is deconvoluted separately from GIT transit, permeation, gut wall

metabolism and first pass liver extraction

P-gp etc

Metabolism

Enterocytes

Faeces

Dissolved Drug

Absorption / Efflux

Degradation

Colon Jejunum I & II Duodenum Ileum I Ileum II Ileum III Ileum IV Stomach

Solid Dosage

PBPK DISTRIBUTION

MODEL LIVER Portal Vein

Dissolution


Mechanistic vs. Conventional IVIVC

CR BCS I OR BCS II Low Extraction CR BCS I OR BCS II High Extraction

CR BCS III OR BCS IV Low Extraction CR BCS III OR BCS IV High Extraction

Differing Permeation/Metabolism in GIT Drugs with Enterohepatic Recirculation

In vivo In vitro Conventional Deconvolution




In vivo In vitro Conventional Deconvolution In vivo In vitro Conventional

Deconvolution

Dissolution Permeation Systemic Input


• Metoprolol is a BCS Class I (High Solubility High Permeability) High First Pass Extraction Drug

ER BCS I OR BCS II High Extraction


Conventional methods may require IVIVR function with lag time, sigmoid or power function to accommodate delay and slowness of in vivo availability as compared to in vitro conditions

• Relatively short half life and sufficient absorption from colon makes it suitable candidate for extended release formulation

• ER formulations of BCS Class I drug are Dissolution/Release Controlled Absorption

Mechanistic IVIVC Case Study Using Simcyp


Reported Model 1: Functional Numerical Deconvolution with Oral Solution as UIR

• The IVIVR was sigmoid or with lag time. • FRA -> FRD graphs shown here used all three formulation data to establish IVIVC. No

graphs are shown when one of the formulations is used as external

In vitro In vivo

In vitro in vivo relationship

USP II, pH 6.8, 50 RPM USP II, pH 6.8, 150 RPM

Eddington et al. (1998) Pharm Res 15(3), 466

Mechanistic IVIVC Case Study: Metoprolol IVIVC


Reported Model 2: Semi Mechanistic Parent/Metabolite Model

USP II, pH 6.8, 150 RPM In vivo Parent and Metabolite Data

In vitro in vivo relationship

Sirisuth and Eddington (2002) EJPS 53, 301

In spite of using parent and metabolite data and lag time IVIVR, prediction errors were higher. No graphs or results are shown when one of the formulations is used as external



Slow and Fast Formulations Used for IVIVC and Medium as External

Deconvolution of in vivo dissolution rather than input rate allowed the establishment of IVIV Relationship without lag time and superior correlation (R2=0.98).

y = 0.8684x R² = 0.9675

y = 0.0021x2 + 0.6925x R² = 0.9782

0

20

40

60

80

100

120

0 50 100 150

%D

isso

lved

in v

ivo

% Dissolved in vitro

Linear IVIVC without Lag

Decon

Linear (Decon)

Poly. (Decon)

y = 0.9941x R² = 0.9703

-20

0

20

40

60

80

100

120

-20 0 20 40 60 80 100 120

Dec

on

vo

lute

d i

n v

ivo

%

Dis

so

lve

d

IVIVR Predicted in vivo % Dissolved

IVIVR Predictivity

• PK Data after oral solution were used to estimate permeability, distribution and elimination parameters

• in vivo dissolution profiles that produce observed PK profiles of each formulation were estimated



Understanding Bio-relevance of dissolution media

• Using two extremes (fast and slow) to develop model and test the model for medium also conforms to the idea of domain of applicability

0102030405060708090

100

0.0 5.0 10.0 15.0

%D

isso

lved

Time (h)

Fast Release Formulation

In vitro Decon

0

20

40

60

80

100

0.0 5.0 10.0 15.0

Time (h)

Medium Release Formulation

In vitro Decon

0

20

40

60

80

100

0.0 5.0 10.0 15.0

Time (h)

Slow Release Formulation

In vitro Decon

• The in vitro dissolution media used is more bio-relevant to medium release formulation but not for the fast and slow release formulations

• Thus, Medium and Fast/Slow release formulations based IVIVC will lead to very good statistics (R2) however confidence in prediction for new formulation beyond used range is questionable



Mechanistic IVIVC Using Simcyp: Internal and External Validations

0

20

40

60

80

100

120

140

-1 4 9 14 19 24

Pla

sm

a C

on

c (

ng

/mL

)

Time (h)

Fast Formulation - Internal

Cobs

Cpred

0

10

20

30

40

50

60

70

80

90

100

-1 4 9 14 19 24Time (h)

Medium Formulation - External

Cobs

Cpred

0

10

20

30

40

50

60

70

80

-1 4 9 14 19 24Time (h)

Slow Formulation - Internal

Cobs

Cpred

Formulation ID Cmax (ng/mL) AUC (ng/mL*h)

Obs Pred %PE Obs Pred %PE

Internal Validation

Fast 108.00 109.61 -1.49 791.00 805.99 -1.89

Slow 63.70 69.54 -9.16 696.00 710.04 -2.02

Mean -5.33 -1.96

External Validation

Medium 86.30 83.51 3.24 812.00 780.50 3.88



Designing a new once-daily formulation of Metoprolol

Fast Formulation Medium

Formulation

Slow

Formulation

Time (h) Cp (ng/mL) Cp Cp

0 0 0 0 1.0 39.61 26.16 18.55 1.5 69.30 46.13 31.71 2.0 92.84 60.91 40.50 3.0 118.68 87.56 56.94 4.0 107.25 90.39 61.71 6.0 69.16 78.23 63.00 8.0 45.98 59.26 53.11

10.0 28.14 39.16 36.24 12.0 18.08 26.83 26.99 14.0 11.91 17.91 17.26 16.0 8.66 13.20 11.58 20.0 4.27 7.19 5.89 24.0 2.79 3.76 2.95

How can you use Simcyp to design new once-daily (24 h)

formulation?

Therapeutic Cp range of metoprolol is 20-100 ng/mL

• All three formulations lead to Cp below 20 ng/mL after 12 h

• Thus they require administration every 12 h (twice daily)

• The Fast formulation leads to Cp beyond maximum tolerable Cp (100 ng/mL)



Designing a new once-daily formulation: Generating desired Cp Profile

Fast Medium Slow New

Time (h) Cp (ng/mL) Cp Cp Cp

0 0 0 0 0.00 1.0 39.61 26.16 18.55 25.00 1.5 69.30 46.13 31.71 34.00 2.0 92.84 60.91 40.50 42.00 3.0 118.68 87.56 56.94 54.00 4.0 107.25 90.39 61.71 62.00 6.0 69.16 78.23 63.00 70.00 8.0 45.98 59.26 53.11 72.00

10.0 28.14 39.16 36.24 70.00 12.0 18.08 26.83 26.99 65.00 14.0 11.91 17.91 17.26 56.00 16.0 8.66 13.20 11.58 48.00 20.0 4.27 7.19 5.89 34.00 24.0 2.79 3.76 2.95 24.00

How to decide the dose???

• Intuitively design a desired PK profile keeping in mind the disposition kinetics of the drug

• Calculate AUC of the known formulation and newly designed PK profile

• Now estimate required dose assuming linear dose-proportionality

Estimated Dose of New Formulation is 200 mg



Designing a new once-daily formulation: Optimising dose to get desired Cp Profile

How much is the desired dose???

• Use disposition parameters obtained previously in IVIVC exercise and estimate required dissolution profile to obtain desired Cp profile

• The best fit was able to achieve desired absorption phase but not the disposition

• This indicates that amount of drug is not sufficient to maintain desired Cp until 24 h

• More dose should be administered to achieve desired total exposure

0.000

10.000

20.000

30.000

40.000

50.000

60.000

70.000

80.000

0.0 5.0 10.0 15.0 20.0 25.0 30.0

Su

b P

lasm

a (

(ng

/mL

))

Time (Hours)

'Subject: 1': Predicted and observed values

Observe IPRED



Designing a new formulation : Optimise the desired dose to achieve desired Cp profile

0.000

10.000

20.000

30.000

40.000

50.000

60.000

70.000

80.000

0.0 5.0 10.0 15.0 20.0 25.0 30.0

Su

b P

lasm

a (

(ng

/mL

))

Time (Hours)

'Subject: 1': Predicted and observed values

Observe IPRED

Formulation AUC Dose No 0 0

Solution 346 50 Fast 791 100

Medium 812 100 Slow 696 100 New 1181 250

0

100

200

300

0 500 1000 1500D

ose (

mg

)

AUC

Estimated Dose

The dose - AUC relationship is not linear

The dose required to achieve the desired Cp profile is 250 mg Now predict the required in vivo dissolution to achieve desired Cp profile



0.000

10.000

20.000

30.000

40.000

50.000

60.000

70.000

80.000

0.0 5.0 10.0 15.0 20.0 25.0 30.0

Su

b P

lasm

a (

(ng

/mL

))

Time (Hours)

Desired and Predicted Cp Profiles

Observe IPRED

0.00

20.00

40.00

60.00

80.00

100.00

120.00

0 4 8 12 16 20 24

Dis

so

luti

on

Mean

(%

)

Time (h)

Required in vivo dissolution Profile

Time %Dissolved 0 0.00 1 6.29 2 8.98 4 16.31 6 23.33 8 30.40 12 46.36 16 63.69 24 99.89

• This is the required in vivo dissolution profile

• Use the established and validated IVIVC to obtain required in vitro dissolution profile from estimated in vivo dissolution profile

Designing a new formulation: Estimate required dissolution profile to achieve desired Cp profile



Designing a new formulation: Assessing the population variability of designed formulation

In none of the trials, Cp has exceeded MTC (100 ng/mL) and maintained Cp of more than MEC (20 ng/mL)

0

20

40

60

80

100

120

0 6 12 18 24 30

Syste

mic

Co

nc

en

trati

on

(n

g/m

L)

Time (h)

Mean Values of Systemic concentration in plasma of Sim-Metoprolol over Time

CSys - Trial 1 CSys - Trial 2 CSys - Trial 3 CSys - Trial 4

CSys - Trial 5 CSys - Trial 6 CSys - Trial 7 CSys - Trial 8

CSys - Trial 9 CSys - Trial 10 CSys Mean

• Simulation trials 10X10 with healthy volunteers aged 20-50



Designing a new formulation: Assessing the steady state performance of designed formulation

At steady state, Cp level was maintained within the therapeutic range of metoprolol.

• Simulated PK profile in healthy volunteers aged 20-50 for 7 days

67.93

26.84

79.73

25.16

79.07

27.96

80.90

28.54

77.86

27.23

81.15

0.00E+00

1.00E+01

2.00E+01

3.00E+01

4.00E+01

5.00E+01

6.00E+01

7.00E+01

8.00E+01

9.00E+01

0 24 48 72 96 120 144

Syste

mic

Co

ncen

trati

on

(n

g/m

L)

Time (h)

Mean Values of Systemic concentration in plasma of Sim-Metoprolol over Time

CSys



• Simcyp can also help to evaluate performance of this new formulation in elderly, paediatric and specific age groups

• Performance on various disease groups, ethnicity, etc. can be studied using Simcyp

• Metabolic DDIs could be evaluated for the designed formulation

• The simulated PK profiles could be linked to pharmacodynamics models to estimate and understand the efficacy of the designed formulation



THANK YOU


Differing Permeation/Metabolism

In vitro

Conventional Deconvolution

Advantages of Mechanistic IVIVC

Mechanistic Deconvolution

In vitro

In v

ivo

Simple IVIVC Function

In v

ivo

Complex IVIVC Function

In vitro

Simple IVIVRs are

important during

formulation optimisation


IVIVC and BCS Classes

BCS

Class

Solubility Permeability Controlling

Factor

IVIVC Expectation for IR

Formulations

Class I High High Gastric Emptying IVIVC expected, if dissolution rate is slower than gastric emptying rate

Class II Low High Dissolution/Sol IVIVC expected

Class III High Low Permeability Permeability is rate determining and limited or no IVIVC with dissolution

Class IV Low Low Diss/Sol/Perm Limited or no IVIVC is expected

Chilukuri et al. (2007) Pharmaceutical Product Development: In vitro in vivo correlation

IVIVC is expected and is used to obtain ‘Biowaiver’ for CR/ER products of Class I drugs as the release/dissolution becomes the rate limiting step

Mechanistic IVIVC models are useful for class III/IV drugs as they distinguish dissolution, GI transit and permeation processes and avoids confounding

IVIVC for drugs with nonlinearity in Fg, Fa and systemic clearance can not be modelled using conventional methods but can be effectively modelled using mechanistic models


Why IVIVC?

• Current Simcyp Implementation: Predict PK profile of Oral formulation from in vitro dissolution, Peff and disposition parameters

Dissolution Time 0 1 2 3 4 5

%Diss 0 25 50 75 100 100

Interpolation

t

%D

ADAM + PBPK

Parameters

t

Cp

Assumption: in vitro dissolution == in vivo dissolution

In vitro Predicted

• In vitro dissolution profiles vary depending on the media, type of formulation and

other experimental conditions used.

• If validated relationship between in vitro and in vivo dissolution (IVIVC) is

obtained, it could be used to effectively predict PK(/PD) profiles new formulations


Why IVIVC

Dissolution Time 0 1 2 3 4 5

%Diss 0 25 50 75 100 100

Interpolation

t

ADAM + PBPK

Parameters

t

Cp

In vitro Predicted

t

%D

In vivo

IVIVC

What IVIVC Relationship to use and How to Establish/Validate it?


IVIVC and Its Components

0

50

100

150

-6 4 14 24

Pla

sm

a C

on

c

(ng

/mL

)

Time (h)

Formulation PK

0

20

40

60

80

100

0.0 4.0 8.0 12.0

%D

isso

lved

/Ab

so

rbe

d

Time (h)


020406080

100

0.0 4.0 8.0 12.0

%D

isso

lved

Time (h)

In vitro Dissolution

y = 0.9941x R² = 0.9703

0

50

100

0 50 100

in v

ivo

%

Dis

so

lved

/Ab

so

rbed

in vitro % Dissolved

IVIVR

IVIVC: A predictive mathematical model describing the relationship between an in vitro property (usually drug dissolution or release) and a relevant in vivo response, e.g., amount of drug absorbed/dissolved.

DECONVOLUTION

IVIVC: The in vitro dissolution and in vivo input curves may be directly superimposable or may be made to be superimposable by the use of a scaling factor/function. Nonlinear correlations, while uncommon, may also be appropriate.

IVIVC

MODELS

MODELS

0

20

40

60

80

100

0.0 4.0 8.0 12.0Time (h)

IVIVR Predicted in vivo Abs

0

50

100

150

-6 4 14 24Time (h)

Predicted PK CONVOLUTION

Validation: External validation is recommended when only 2 formulations are used to establish IVIVC

VALIDATION


Simcyp IVIVC – Two Stage

%D

STAGE 1

ADAM + PBPK

Parameters

t

Cp

In vivo

%D

In vitro Input

Discrete Data

Time 0 1 2 3 4 5

%Diss 0 25 50 75 100 100

Vivo

%D

IVIVC

Vitro %D

Observed

STAGE 2

Establish IVIVC

In vitro


Single Stage IVIVC

In vitro

Dissolution

Time 0 1 2 3 4 5

%Diss 0 25 50 75 100 100

Interpolation

ADAM + PBPK

Parameters

In vitro

IVIVC

Predicted

In vivo IVIVC Function

And Initial

Parameters

IVIVC

Predicted

Observed

Iteratively estimate IVIVC Function Parameters

Which give best fit of predicted PK profile

to the observed PK Profile

mechanistic ivivc using the simcyp adam model -...

Documents