QbD in Pharmaceutical Development: Tablet formulation design … - QbD in... · 2011. 7. 12. · Wet Direct Roller Granulation Compression Compaction Unit operations, no. 6 3 5 Scale

QbD

in P

harm

aceu

tical

Dev

elop

men

t:T

able

t for

mul

atio

n de

sign

spa

ces

for

dire

ct

com

pres

sion

and

rol

ler

com

pact

ion

proc

esse

s

Mor

ten

Alle

sø,

PhD

(M

Sc

Pha

rm)

Pha

rmac

eutic

al s

cien

tist

ISP

E N

ordi

c PA

T C

oP, J

une

1st

2011

KU

-LIF

E, C

open

hage

n

Out

line

1.

In

tro

du

cti

on

to

tab

let

form

ula

tio

ns a

nd

ch

all

en

ges (

CQ

As, excip

ien

ts, C

U

an

d d

isso

luti

on

)2

.O

vervie

w o

f p

ro

cesses a

nd

un

it o

perati

on

s –

ch

osin

g t

he m

ost

co

st

eff

ecti

ve

man

ufa

ctu

rin

g3

.W

hat

is a

sta

nd

ard

fo

rm

ula

tio

n?

4.

Dir

ect

co

mp

ressio

n1.

Composition

2.

Dose homogeneity and theoretical considerations (incl. comparison to practice)

3.

NIR blend data

4.

Design space construction

2

4.

Design space construction

5.

Conclusion and perspectives (more DC form

. implementation i development through decision-

making trees)

5.

Ro

ller c

om

pacti

on

(d

ry g

ran

ula

tio

n)

1.

Introduction to roller compaction

2.

Critical quality attributes of ribbons and the granule (porosity, fines and API in fines)

3.

Ranking of critical process variables

4.

Design space of ribbon and granule properties

5.

Future work

6.

Co

nclu

din

g r

em

arks

Tabl

et fo

rmul

atio

ns a

nd m

anuf

actu

ring.

..

3

Tabl

et fo

rmul

atio

ns a

nd m

anuf

actu

ring.

..

Com

mon

tabl

et c

ompo

sitio

ns

Act

ive

ingr

edie

nt (

AP

I)F

iller

(s)

(+bi

nder

)Lu

bric

ant

Low

dos

e (0

.5%

AP

I)In

term

edia

te d

ose

(10%

AP

I)

Oth

er (

not s

how

n)

4

Hig

h do

se (

20%

AP

I)V

ery

high

dos

e (5

0% A

PI)

Tab

let p

roce

ss te

chno

logi

es

Wet

gra

nula

tion

high

she

ar m

ixer

Wet

Gra

nula

tion

Dire

ct C

ompr

essi

on

Mix

ing

of A

PI w

ithex

cipi

ents

Rol

ler C

ompa

ctio

n

Wet

sie

ving

Dry

ing

in fl

uid

bed

Rol

ler c

ompa

ctio

n

Mix

ing

of A

PI w

ithex

cipi

ents

Tab

let c

ompr

essi

on

Dry

sie

ving

Dry

sie

ving

5

Dry

ing

in fl

uid

bed

Dry

sie

ving

Mix

ing

gran

ulat

e w

ithex

cipi

ents

Tab

let

com

pres

sion

Mix

ing

gran

ule

with

lubr

ican

t

Tab

let c

ompr

essi

on

Tab

let

com

pres

sion

Low

dos

e (0

.5%

AP

I)

Cha

lleng

e: D

ose

hom

ogen

eity

�P

artic

le s

ize

of A

PI i

s cr

itica

l!�

AP

I par

ticle

siz

e co

ntro

l and

the

blen

ding

ste

p ar

e ve

ry c

ritic

al

Inte

rmed

iate

dos

e (1

0% A

PI)

6

Ver

y hi

gh d

ose

(50%

AP

I)C

halle

nge:

Mak

ing

the

tabl

et s

tick

�R

equi

res

good

com

pres

sion

pro

pert

ies

of

AP

I!�

Dry

pro

cess

ing

met

ho

ds

usu

ally

no

t ap

plic

able

Inte

rmed

iate

dos

e (1

0% A

PI)

Hig

h do

se (

20%

AP

I)

Usu

ally

wel

l sui

ted

for

dire

ct

com

pres

sion

and

rol

ler

com

pact

ion

Rat

iona

l sel

ectio

n of

pro

cess

tech

nolo

gy

Exa

mpl

e:IR

-tab

let,

3 st

reng

ths,

10,

20

and

30 m

gD

ose

prop

ortio

nal f

orm

ulat

ion

Par

amet

erW

etD

irect

R

olle

rG

ranu

latio

nC

ompr

essi

onC

ompa

ctio

nU

nit o

pera

tions

, no.

63

5S

cale

up

com

plex

sim

ple

sim

ple

QbD

com

plex

sim

ple

sim

ple

7

QbD

com

plex

sim

ple

sim

ple

AP

I- P

SD

crit

ical

itylo

wer

high

erlo

wer

AP

I sup

ply

until

sca

le-u

p to

pro

duct

ion,

kg

9035

35P

rodu

ctio

n cy

cle

time,

inde

x30

050

100

Typ

ical

pro

duct

ion

thro

ughp

ut, k

g/h

125

400

400

Pre

ferr

ed (

if ap

plic

able

)

12

The

sta

ndar

d fo

rmul

atio

n co

ncep

t...

8

The

sta

ndar

d fo

rmul

atio

n co

ncep

t...

Wha

t is

a st

anda

rd fo

rmul

atio

n?

Sta

nd

ard

fo

rm

ula

tio

n (

SF):

�A

fo

rm

ula

tio

n t

em

pla

te a

pp

lied

to

new

(an

d s

uit

ab

le) H

Lu

AP

Is.

Th

oro

ug

h a

nd

str

uctu

red

do

cu

men

tati

on

is a

vailab

le f

or e

ach

sta

nd

ard

fo

rm

ula

tio

n,

describ

ing

th

e r

ela

tio

nsh

ip b

etw

een

crit

ical

form

ula

tio

n-

an

d p

ro

cess v

aria

ble

s a

nd

on

e o

r s

everal

crit

ical

qu

ality

att

rib

ute

s

Des

ign

spac

e

9

Ad

van

tag

es o

f a S

F:

�In

-dep

th k

no

wle

dg

e o

f th

e f

orm

ula

tio

n i

s a

lread

y a

vailab

le,

i.e.

develo

pab

ilit

yw

ell-k

no

wn

.

�R

ed

ucti

on

of

wo

rklo

ad

of

a d

evelo

pm

en

t p

ro

ject

(efficiency)

Des

ign

spac

e

A s

tand

ard

form

ulat

ion

is a

pla

tform

tech

nolo

gy

(in

-ho

use

term

)

Des

ign

spac

e de

finiti

on

�D

esig

n s

pace:

IC

H Q

8: ”The multidimensional combination and

interaction of input variables(e.g., material

attributes) and process parametersthat have been

demonstrated to provide assurance of quality.

Working within the design space is not considered as

Mul

ti-fa

ctor

ial,

DoE

10

Working within the design space is not considered as

a change”

Uni

t ope

ratio

nsC

ritic

al p

roce

ss p

aram

eter

s (C

PP

)

The

form

ulat

ion,

raw

mat

eria

ls

Crit

ical

qua

lity

attr

ibut

es (

CQ

A)

-rel

ated

to s

peci

ficat

ions

A p

rere

quis

ite fo

r th

e co

nstr

uctio

n of

a d

esig

n sp

ace

is D

oE!

(impl

emen

ting

a co

ntr

ol s

trat

egy

usua

lly r

equi

res

PAT

)

Crit

ical

Qua

lity

Attr

ibut

es o

f the

fina

l tab

lets

Ble

ndin

g tim

e an

d lu

bric

atio

n bl

edin

g tim

e (e

stab

lish

ed b

ased

on

prel

imin

ary

expe

rimen

ts)

-(C

PP

-C

ritic

al P

roce

ss P

aram

eter

)

Tabl

etin

g pr

oper

ties

–Com

pact

ibili

ty a

nd c

ompr

essi

bilit

y–A

dhes

ion

to p

unch

esTa

blet

tech

nica

l pro

pert

ies

–Har

dnes

s >

50

N

11

–Har

dnes

s >

50

N–F

riabi

lity

< 0

.5%

–Dis

inte

grat

ion

time

< 5

min

.

Con

tent

uni

form

ity (

RS

D)

≤ ≤≤≤4%

(C

QA

)

Seg

rega

tion

prop

ertie

s (e

.g. f

luid

izat

ion)

(C

QA

)

Dis

solu

tion?

Sta

bilit

y?A

PI d

epen

dent

(m

olec

ular

pro

pert

ies)

!

Dire

ct c

ompr

essi

on fo

rmul

atio

n...

12

Dire

ct c

ompr

essi

on fo

rmul

atio

n...

Cho

ice

of e

xcip

ient

s-

prev

ious

kno

wle

dge

/ lite

ratu

re

�Fille

r:

Mic

ro

crysta

llin

e c

ellu

lose

× ×××Compatible with m

any APIs

× ×××Plastic deform

ing m

aterial (excellent compactibility)

× ×××Multi-functional: Disintegrant+binder

× ×××Relatively good flowability

�(S

up

er) d

isin

teg

ran

t: C

ro

scarm

ell

ose s

od

ium

× ×××Cross-bound carboxymetylcellulose

13

× ×××Cross-bound carboxymetylcellulose

× ×××Low conc. required

× ×××Effective disintegration agent

× ×××May be incompatible w/ API (ionin interaction)

�Lu

bric

an

t: M

g-s

tearate

× ×××Most effective (anti-adhesive effect)

× ×××Disrupts internal tablet bondings (affects dissolution/disintegration)

× ×××May be incompatible w/ API

For

mul

atio

n co

mpo

sitio

ns a

nd P

SD

of A

PI’s

(es

cita

lopr

am)

�A

PI (

1,

11

an

d 2

0%

w/

w)

�A

vic

el P

H1

02

(d

50

= 1

22

µm

, d

90/

d5

0=

2.0

)

�3

% A

c-d

i-so

l (cro

scarm

ello

se s

od

ium

)

�1

% o

r 2

% M

g-s

tearate

(2

% a

t 2

0%

AP

I)

�A

PI o

f q

ualiti

es D

50=

42

µm

, 6

7 µ

m,

11

4 µ

m a

nd

16

3

µm

(la

ser d

iffr

acti

on

)

14

AP

Id 1

0, µ

md 5

0, µ

md 9

0,µ

md 5

0/d 1

0d 9

0/d50

16.

242

165

6.8

42

8.6

6721

17.

83.

23

1111

424

910

2.2

418

163

338

9.1

2.1

Cen

tral

Com

posi

te F

ace

(CC

F)

desi

gn42

42

µ µµµµ µµµm

AP

I qua

lity

was

add

edm

AP

I qua

lity

was

add

edto

the

desi

gn w

ithou

t to

the

desi

gn w

ithou

t co

mpr

omis

ing

the

desi

gn

com

prom

isin

g th

e de

sign

qu

ality

qual

ity

The

aim

with

the

desi

gned

exp

erim

ent i

s to

app

roxi

mat

e th

e re

spon

se (

CU

-rsd

) by

a q

uadr

atic

pol

ynom

ial (

mod

el)

in o

rder

to:

•Und

erst

and

in m

ore

deta

il ho

w fa

ctor

s: p

artic

le s

ize

and

drug

load

influ

ence

the

resp

onse

; i.e

. get

a m

ap o

f the

sys

tem

(re

spon

se s

urfa

ce m

odel

ling)

.•M

ake

pred

ictio

ns a

nd fi

nd a

reg

ion

of o

pera

bilit

y.T

hus,

the

desi

gn w

ill a

lso

inve

stig

ate

for

any

inte

ract

ion

effe

cts

and

non-

linea

ritie

s in

th

e da

tase

t.

15

the

data

set.

Ana

lyze

2 ta

blet

s fo

r co

nten

t for

eve

ry 2

000t

h pr

oduc

ed (

100

min

tabl

etin

g)C

alcu

late

RS

D o

ver

tabl

etin

g =

Dos

e ho

mog

enei

ty

Ble

ndin

g ev

alua

tion

by N

IR-S

entr

oPat

16

Sen

troP

AT

NIR

spe

ctro

met

er

(Sen

tron

ic G

mbH

, Dre

sden

, Ger

man

y)

The

oret

ical

opt

imal

dos

e ho

mog

enei

ty (

Sre

l = C

U%

)

()

WP

wP

wP

PS

A

bA

aB

Bre

lA

+=

100

%,

P A/B

: rel

ativ

e co

nten

t of A

PI a

nd A

vice

l PH

102,

res

pect

. (P A

+PB=1

)w

a: m

ean

wei

ght (

base

d on

vol

ume

wei

gthe

d vo

lum

e m

ean

∑=

ii

ia

dv

w3

6ρ

π

17

wa

: mea

n w

eigh

t (ba

sed

on v

olum

e w

eigt

hed

volu

me

mea

n

diam

eter

) of

a s

ingl

e pa

rtic

le o

f AP

I and

Avi

cel P

H10

2,

resp

ectiv

ely,

cal

cula

ted

from

the

volu

me

size

dis

trib

utio

n (la

ser

diffr

actio

n)

Assu

mes f

ree f

low

ing

no

n-i

nte

racti

ng

bin

ary m

ixtu

res!

In

th

is c

ase:

ρ ρρρ AP

I≈ ≈≈≈

ρ ρρρ Avic

el

PH

10

2(red

uces r

isk o

f seg

reg

ati

on

)

The

oret

ical

opt

imal

dos

e ho

mog

enei

ty v

ersu

s ex

perim

enta

l

1 m

g ta

ble

t

4.0

6.0

8.0

10.0

12.0

Content RSD (%)

Act

ual

Theo

retic

al

1 m

g ta

blet

1 m

g ta

blet

20 m

g ta

blet

20 m

g ta

blet

20 m

g ta

ble

t

3.0

4.0

5.0

6.0

Content RSD (%)

Act

ual

Theo

retic

al

20 m

g ta

blet

20 m

g ta

blet

18

0.0

2.0

4267

114

163

50 m

g ta

ble

t

0.0

1.0

2.0

3.0

4.0

5.0

6.0

42**

6711

416

3

Content RSD (%)

Act

ual

Theo

retic

al

50 m

g ta

blet

50 m

g ta

blet

Adh

esio

n to

pun

ches

!A

dhes

ion

to p

unch

es!

0.0

1.0

2.0

42**

6711

411

416

3Content RSD (%)

Con

stru

ctin

g th

e de

sign

spa

ce:

Mul

tiple

Lin

ear

Reg

ress

ion

(MLR

)

The

dat

a fr

om th

e st

atis

tical

exp

erim

enta

l des

ign

can

be fi

tted

by m

eans

of m

ultip

le li

near

reg

ress

ion

(MLR

) an

d th

e re

spon

se, y

, can

be

desc

ribed

by

a po

lyno

mia

l fu

nctio

n:C

UR

SD

= b

1k1

+ b

2k2

+ in

tera

ctio

n te

rms

+ c

onst

ant

19

Des

ign

spac

e vi

sual

izat

ion

base

d on

MLR

API, % of tablet

20

AP

I

API, % of tablet

Con

clus

ions

for

DC

sta

ndar

d fo

rmul

atio

n ba

sed

on A

vice

l PH

102

�T

heo

reti

cal

calc

ula

tio

n o

n o

pti

mal

do

se h

om

og

en

eit

y c

orrela

tes

well

wit

h t

he a

ctu

al

valu

es a

s a

very v

alu

ab

le t

oo

l fo

r p

red

icti

on

of

the e

ffect

of

AP

I-P

SD

on

CU

% f

or t

he f

inal

tab

lets

based

on

DC

�A

desig

n s

pace w

as e

sta

blish

ed

base o

n C

FF d

esig

n s

ho

win

g t

he

eff

ect

of

AP

I o

n d

ose h

om

eg

en

eit

y i

n t

he d

ose r

an

ge o

f 1

-2

0%

A

PI.

�B

elo

w A

PI-P

SD

of

d5

0<

ap

pro

x.

40

µm

ad

hesio

n t

o p

un

ch

es o

ccu

r

for 2

0%

AP

I e

ven

at

2%

Mg

-ste

arate

→ →→→E

xte

rn

al

lub

ric

ati

on

21

for 2

0%

AP

I e

ven

at

2%

Mg

-ste

arate

→ →→→E

xte

rn

al

lub

ric

ati

on

�A

vic

el

PH

10

2 i

s a

n e

ffecti

ve f

ille

r/

dis

inte

gran

t p

ro

vid

ing

excellen

t

co

mp

acti

bilit

y a

nd

co

mp

ressib

ilit

y a

nd

tab

let

dis

inte

grati

on

tim

es

< 2

0 s

ec w

ith

3%

su

perd

isin

teg

ran

t A

c-D

i-S

ol.

Ac-D

i-S

ol

mig

ht

be

exclu

ded

(to

be i

nvesti

gate

d)

�S

eg

reg

ati

on

du

e t

o f

luid

izati

on

is r

ed

uced

wh

en

th

e A

PI-P

SD

an

d

do

se i

s r

ed

uced

(d

ue t

o i

ncreased

co

hesio

n b

etw

een

parti

cle

s)

On

goin

g w

ork

on D

C s

tand

ard

form

ulat

ions

�T

he u

se o

f u

sin

g e

xte

rn

al

lub

ric

ati

on

wit

h i

n-p

ro

cess s

prayin

g o

f m

ag

nesiu

m s

tearate

to

elim

inate

po

ten

tial

inco

mp

ati

bil

ity b

etw

een

M

g-s

tearate

an

d A

PI.

�A

ltern

ati

ve f

ille

rs t

han

MC

C i

s i

nvesti

gate

d.

Fo

rm

ula

tio

n s

creen

ing

h

as s

ho

wn

th

e b

est

can

did

ate

s t

o b

e:

Flo

wla

c 1

00

(la

cto

se) a

nd

P

arte

ck M

20

0 (

man

nit

ol)

. A

bin

der i

s n

eed

ed

to

im

pro

ve

co

mp

acti

bilit

y o

f th

e f

inal

ble

nd

.

22

co

mp

acti

bilit

y o

f th

e f

inal

ble

nd

.

Rol

ler

com

pact

ion

form

ulat

ion.

..

23

Rol

ler

com

pact

ion

form

ulat

ion.

..

Rol

ler

com

pact

ion

24

Why

do

dry

gran

ulat

ion

Ad

van

tag

es:

•Im

pro

ve f

low

ab

ilit

y

•Im

pro

ve w

eig

ht

un

ifo

rm

ity

•Im

pro

ve c

on

ten

t u

nif

orm

ity (

po

ssib

ly)

•M

inim

ize s

eg

reg

ati

on

(p

ossib

ly)

CQ

As (

gran

ula

te):

Rib

bo

n p

oro

sit

y (

wo

rk h

ard

en

ing

)W

ork

hard

enin

g (t

able

tabi

lity)

25

•R

ibb

on

po

ro

sit

y (

wo

rk h

ard

en

ing

)

•N

o.

of

fin

es

•B

imo

dal /

un

imo

dal?

•D

istr

ibu

tio

n o

f A

PI i

n s

ieve

fracti

on

s

Wor

k ha

rden

ing

(tab

leta

bilit

y)

Det

erm

ines

impa

ct o

f se

greg

atio

n on

tabl

et c

onte

nt

unifo

rmity

CP

Ps (

RC

):

•C

om

pacti

on

fo

rce

•G

ap

siz

e

•R

oll s

peed

(m

ostl

y u

p-s

calin

g)

•S

ieve s

ize (

tab

let

do

se u

nif

orm

ity)

Ran

king

of c

ritic

al p

roce

ss p

aram

eter

s (C

PP

s)

C

ritic

al Q

ualit

y A

ttrib

ute

(CQ

A)

Pro

cess

/ fo

rmul

atio

n pa

ram

eter

Rib

bon

poro

sity

F

ines

frac

tion

(< 1

25 µ µµµm

) A

PI-

% in

fin

es fr

actio

n

Com

pact

ion

forc

e (k

N)

HIG

H

INT

ER

ME

DIA

TE

U

NK

NO

WN

Gap

siz

e (m

m)

INT

ER

ME

DIA

TE

U

NK

NO

WN

U

NK

NO

WN

26

Gap

siz

e (m

m)

INT

ER

ME

DIA

TE

U

NK

NO

WN

U

NK

NO

WN

Rol

l spe

ed (

rpm

) LO

W

LOW

(up

on

decr

ease

in r

oll

spee

d)

UN

KN

OW

N

Rol

l wid

th (

cm)

NO

NE

N

ON

E

NO

NE

Gra

nula

tor

angl

e (° °°°)

N

ON

E

LOW

U

NK

NO

WN

Gra

nula

tor

spee

d (r

pm)

NO

NE

N

ON

E

NO

NE

Scr

een

type

N

ON

E

HIG

H

UN

KN

OW

N

Scr

een

size

N

ON

E

INT

ER

ME

DIA

TE

IN

TE

RM

ED

IAT

E

CQ

A s

peci

ficat

ions

of i

nter

med

iate

pro

duct

•Rib

bon

poro

sity

∈[3

5 ; 4

0%]

•% r

elat

ive

cont

ent o

f AP

I in

fines

frac

tions

sha

ll be

min

imiz

ed (

≤45

%)

•Fin

es fr

actio

n sh

all b

e m

inim

ized

(≤

30%

)

Por

osity

det

erm

inat

ion:

Ass

esse

d by

oil-

intr

usio

n m

etho

d (in

-hou

se)

27

Ass

esse

d by

oil-

intr

usio

n m

etho

d (in

-hou

se)

Sev

eral

oth

er m

etho

ds w

ere

initi

ally

com

pare

d

Fin

es fr

actio

n of

gra

nula

te:

Det

erm

ined

by

siev

e an

alys

is

Rel

ativ

e A

PI c

onte

nt in

fine

s fr

actio

n:D

eter

min

ed b

y U

V s

pect

roph

otom

etry

Cal

cula

ted

as %

of t

otal

dos

e re

cove

red

in a

giv

en g

ranu

le fr

actio

n

NO

TE

: A

PI p

artic

le s

ize

(d5o

fixed

at 6

5 µm

)!

Cen

tral

com

posi

te fa

ce d

esig

n

3 f

acto

rs

�Drug load (% w/w

)

�Compaction force (kN)

�Gap size (mm)

2 c

en

ter p

oin

ts

28

2 c

en

ter p

oin

ts

Load

(%

) C

omp.

forc

e (k

N)

G

ap (

mm

)

20

3 2

20

9 2

20

3 4

20

9 4

20

6 3

1.1

3 2

1.1

9 2

1.1

3 4

1.1

9 4

1.1

6 3

11

3 3

11

9 3

11

6 2

11

6 4

11 (

CP

) 6

3 11

(C

P)

6 3

Res

ults

from

rol

ler

com

pact

ion

tria

ls

Por

osity

-%

Fin

es-%

R

el. A

PI c

onte

nt (

%)

Load

(%

) C

omp.

fo

rce

(kN

)

Gap

(m

m)

Thr

ough

put

(kg

/hr)

Ela

stic

re

cove

ry

(mm

) d

gw (

µ µµµm)

Thr

ough

put

Oil

< 71

µ µµµm

<

125

µ µµµm

AP

I-R

SD

in

siev

e fr

ac. (

%)

<

71 µ µµµ

m

< 12

5 µ µµµm

1.1

3 2

6.9

0.2

193

45.1

43

.4

12.3

30

.1

76.6

35

.1

55.2

1.

1 9

2 9.

4 0.

2 31

4 25

.8

21.3

5.

7 14

.8

41.6

11

.9

20.4

1.

1 3

4 12

.2

0.3

201

49.4

51

10

.8

27.3

65

.6

27.2

51

.2

1.1

9 4

16.3

0.

3 25

7 32

.7

31.2

9.

3 21

.8

34.0

16

.5

26.9

1.

1 6

3 11

.6

0.2

231

36.1

35

.3

9.7

24.3

53

.2

22.9

37

.2

11

3 3

10.0

0.

3 16

6 45

.4

42.8

15

.3

35.3

72

.2

36.1

65

.8

11

9 3

13.8

0.

3 23

9 24

.7

24.3

9.

2 23

.0

50.4

21

.1

36.8

11

6

2 8.

6 0.

2 21

9 30

.6

27.6

9.

4 24

.5

55.7

22

.7

41.9

11

6

4 15

.9

0.3

221

32.7

34

.8

11.0

25

.8

58.0

26

.0

45.9

11

6

3 12

.3

0.3

202

32.4

33

.3

11.3

28

.2

57.2

26

.4

49.0

11

6

3 11

.9

0.3

215

34.8

33

.9

10.7

26

.1

60.4

27

.7

40.5

2911

6 3

11.9

0.

3 21

5 34

.8

33.9

10

.7

26.1

60

.4

27.7

40

.5

20

3 2

7.1

0.2

182

41.9

40

.8

11.9

31

.5

57.3

26

.1

54.3

20

9

2 9.

5 0.

2 21

8 21

.7

19.8

9.

7 27

.7

46.9

29

.3

42.1

20

3

4 13

.4

0.25

1 15

1 38

.3

45.4

13

.6

39.4

68

.1

18.6

66

.0

20

9 4

17.5

0.

3 23

4 24

.7

27.1

7.

8 23

.5

41.5

16

.1

35.2

20

6

3 12

.2

0.3

224

32.1

33

.6

8.9

25.6

46

.6

18.5

41

.4

Des

ign

spac

e co

nstr

uctio

n (P

LS m

odel

ing)

0.6

0.8

1.0

R2

Q2

Mod

el V

alid

ityR

epro

duci

bilit

y

30

0.0

0.2

0.4

0.6

TP-porosity

Oil-Porosity

Effe

cts

of C

PP

s on

rib

bon/

gran

ule

prop

ertie

s

31

Mod

el e

quat

ion:

Rib

bon

poro

sity

424446

424446

424446

32

2022242628303234363840

510

1520

Oil-Porosity

Load

2022242628303234363840

3.0

4.0

5.0

6.0

7.0

8.0

9.0

Com

p. fo

rce

2022242628303234363840

2.0

2.5

3.0

3.5

4.0

Gap

Mod

el e

quat

ion:

Fin

es fr

actio

n (%

< 1

25 µ

m)

323436

323436

33

18202224262830

24

68

1012

1416

1820

Mod

el e

quat

ion:

AP

I in

gran

ulat

e fr

actio

n (<

125

µm

)

556065

556065

34

253035404550

24

68

1012

1416

1820

API125

Load

253035404550

3.0

4.0

5.0

6.0

7.0

8.0

9.0

Com

p. fo

rce

Opt

imal

pro

cess

set

tings

: Hitt

ing

the

swee

t spo

t

35

Sw

eet s

pot p

lot s

how

ing

the

oper

atin

g re

gion

(de

sign

spa

ce)

for

each

dru

g lo

ad. A

red

sta

r in

dica

tes

the

oper

atin

g co

nditi

ons

used

in u

pcom

ing

verif

icat

ion

batc

hes.

RC

: Con

clus

ions

and

futu

re w

ork

�C

QA

s c

ou

ld b

e w

ell m

od

ele

d a

gain

st

dru

g l

oad

an

d C

PP

s:

co

mp

acti

on

fo

rce a

nd

gap

siz

e.

�N

ext

ste

p:

Verif

icati

on

of

sw

eet

sp

ot

sett

ing

s i

n a

new

exp

erim

en

tal serie

s (

on

-go

ing

; la

rg

er s

cale

) → →→→

prelim

inary

resu

lts s

ho

w g

oo

d r

ep

ro

du

cib

ilit

y f

or p

oro

sit

y a

nd

mo

stl

y

for f

ines f

racti

on

, h

ow

ever,

new

ch

allen

ges a

re n

ow

36

for f

ines f

racti

on

, h

ow

ever,

new

ch

allen

ges a

re n

ow

en

co

un

tered

:

Sampling related issues when m

easuring fines (sample divider)

Tableting step: Adhesion to punches ⇔

further decrease amount of

fines in the granulate (varying the sieve size of the granulator)

�Fin

ally:

Desig

n s

pace d

escib

ing

rela

tio

nsh

ip b

etw

een

AP

I

load

an

d A

PI p

arti

cle

siz

e (

sim

ilar a

s f

or D

C a

nd

usin

g

verif

ied

sw

eet

sp

ot

sett

ing

s f

or t

he R

C s

tep

)

Con

clud

ing

rem

arks

...

37

Con

clud

ing

rem

arks

...

�A

SF n

eed

s t

o b

e c

on

tin

uo

usly

mo

dif

ied

an

d r

efi

ned

!

�R

esu

lts/

data

an

d f

inally in

form

ati

on

is u

sele

ss u

nle

ss u

sed

as p

art

of

a s

tru

ctu

red

fram

ew

ork (

Qb

D).

So

far t

here a

re

13

exp

erim

en

tal p

lan

s a

nd

rep

orts

beh

ind

th

e S

F w

ork!

38

�D

ecis

ion

-makin

g t

oo

l in

Develo

pm

en

t: I

nte

grate

SF r

esu

lts

as f

low

ch

arts

an

d i

n r

isk a

ssessm

en

t te

mp

late

s (

on

-go

ing

).

�G

oo

d a

dvic

e:

Do

n’t

un

deresti

mate

th

e u

sefu

lness o

f ’o

ld’

data

an

d p

revio

us s

tud

ies.

Seek in

form

ati

on

an

d d

ecrease

the e

xp

erim

en

tal

wo

rk lo

ad

marked

ly (

less c

om

ple

x D

oEs).

Tha

nk y

ou...

Que

stio

ns?

39

Tha

nk y

ou...

Que

stio

ns?