Dynamic Image Analysis and pharmaceutical multi-particulates · Dynamic Image Analysis and...

Dynamic Image Analysisand pharmaceutical multi-particulates

Particle size distributionsCoat thickness and drug release

Swelling during drug release

G. Heinicke© Horiba Instruments Webinar, June 2009

TopicsWhat are multi-particulates and what is DIA?Representative samplesApplications

Raw materialsDrug layeringMeasuring coat thicknessInvestigating drug release

Future opportunities


Multi-particulatesInert CoreDrug + BinderPolymer CoatingDrug release

Sugar Sphere Plus Drug Layer Plus Polymer Layer

Ghost After Dissolution,

Expansion Noticeable

Sugar Sphere Plus Drug Layer

Sugar Sphere


Standard for beads Common, conceptually easy

Sizes 500,600,710,850,1000,1180,1400 µm

Sieving facts

Tolerances on sieves (e.g. 850 µm)–Mean opening 850 ± 35 µm, NMT 5% > 925 µm, max individual 970 µm

–Sieve-to-sieve differencesInfinite shaking, lengthwise passage


Alternate PSDs

MicroscopesLabor intensive, small sample size, which dimension to measure if irregular

Dynamic Image AnalysisImage dimensions irrespective of orientationAverage is larger, conceptually and measured


Dynamic image analysis

Screen

Light Source

Sample Collection

Vibrating Chute and Sample


Measuring shadows


A typical PSD


Particle sizes and distributions


Raw materials, drug layering


Zoom camera images


DistributionsAbsolute homogeneity – 1 particle is sufficientDegree of heterogeneity

Span from smallest to largest particleShape consistency

Typical multi-particulatesDrug layered seed cores (sugar spheres or MCC)Extruded productsHigh shear granulations

Adequate representation


Representative measurements

Different distributionsSize of sampleMethod of samplingHow representative is the sample?Reproducibility of measurement

Data from a given sample is reproducible

Heinicke & Schwartz, Pharm Dev Tech, 9 (4), 359, 2004


Fluid bed sampling

PSD is steady at about 15,000 particles for layered sugar spheres (10-30 g depending on D50 and density)Sample port equivalent to thiefSample port samples represent the processing

Heinicke & Schwartz, Pharm Dev Tech, 9 (4), 359, 2004


Typical data

Applications


Multi-particulates

Sugar Sphere Plus Drug Layer Plus Polymer Layer

Ghost After Dissolution,

Expansion Noticeable

Sugar Sphere Plus Drug Layer

Sugar Sphere


Drug layering

Heinicke et al., Pharm Dev Tech, 10, 85, 2005


Resolve polymer coat addition


These are the particles

13% 15%


Series of coat thickness


900

910

920

930

940

950

960

970

980

990

D50

Mes

(um

)

0 5 10 15 20TPCW (%)

Linear Fit D50Mes (um) = 891.86135 + 4.7723849 TPCW (%) Summary of Fit RSquare 0.993486RSquare Adj 0.992183Root Mean Square Error 1.777776Mean of Response 955Observations (or Sum Wgts) 7


Drug release

0

20

40

60

80

100

0 240 480 720 960 1200

Time (min)

% R

elea

sed

7%CW9%CW11%CW13%CW15%CW17%CW19%CW


TPCW and drug release

R2 = 0.9986

y = 58.608x - 99.382

0

200

400

600

800

1000

1200

5 10 15 20

TPCW (%)

T50

(min

)


Coat thickness & drug release

y = 23.637x - 86.566R2 = 0.9917

0

200

400

600

800

1000

1200

0 10 20 30 40 50

Coat Thickness (um)

T50

(min

)

T50 (min)Linear (T50 (min))


700

750

800

850

900

950

1000

1050

1100

D50

Mes

0 5 10 15 20TPCW


How real is it?

Supporting evidence for meaningful data


Effect of batch surface area

Adjust by altering core charge weightRatio of weights is ratio of area (different weights of same substrate)Normalize for TPCW and look at drug release

Effect of raw material PSD (SA) on coat thickness, a contribution to lot-to-lot variation


Experiment with batch size

600, 650, 700, 750 and 800 g core chargesOne coating solutionSame coating solution weight additions of 425, 559, 699, 845, 998, 1158, and 1326 g to each core chargeComparison of coat thicknessComparison of drug release


Core charge & change in D50

960

970

980

990

1000

1010

1020

1030

1040

D50

6 8 10 12 14 16 18 20Nominal CW

Linear Fit Bx ID=="600g"Linear Fit Bx ID=="650g"Linear Fit Bx ID=="700g"Linear Fit Bx ID=="750g"Linear Fit Bx ID=="800g"


Normalize to TPCW

960

970

980

990

1000

1010

1020

1030

1040

D50

5 10 15 20TPCW



TPCW and T50

100200300400500600700800900

10001100

T50

(min

)

5 10 15 20TPCW



Explain the single line


Different experiments marked

100200300400500600700800900

10001100

T50

(min

)

5 10 15 20TPCW



Novel use of DIA

small

medium

large Low potency

High potency



Different SS sieve cuts


Expt 1: Coating constant PSDsSame sized particles700 g core charge of each (separately)Same coating solution weight additions of 425, 559, 699, 845, 998, 1158, and 1326 g to each core chargeSample, measure D50 at each TPCW, plotQ: Are the lines coincident, parallel or diverging?


Constant coat thickness

small

medium

large Low potency

High potency


10

15

20

25

30

35

40

45

Coa

tTh

ickn

ess

(um

)

6 8 10 12 14 16 18 20TPCW

Expt 1: Slopes aren’t equal

Linear Fit Core=="15"Linear Fit Core=="45"Linear Fit Core=="64"


DensityD

ensi

ty

0.65

0.7

0.75

0.8

0.85

15% 45% 64%

Potency at Same Size

All PairsTukey-Kramer 0.05



Adjust for surface area

CddC 1

1

2

1

22 ρρ

=

Cx = Core Charge Weight (1 or 2)

ρx = Bulk Density of Cores

dx = Diameter of CoresHeinicke et al., Pharm Dev Tech, 10, 85, 2005


Expt 2: Adjust Core ChargeSame sized particles636g, 700 g, 775 g sprayed separatelySame coating solution weight additions of 425, 559, 699, 845, 998, 1158, and 1326 g to each core chargeSample, measure D50 at each TPCW, plotQ: Are the lines coincident, parallel or diverging?


Expt 2: Slopes are equal

10

15

20

25

30

35

40

Coa

t Thi

ckne

ss (u

m)

6 8 10 12 14 16 18 20 22TPCW (%)

Linear Fit Core ID=="15"Linear Fit Core ID=="45"Linear Fit Core ID=="64"


Investigating drug releasePolymer insolubleSwelling and stretchingCoat thickness effectsCore dependent


Coat thickness and swelling


Constant coat thickness

small

medium

large Low potency

High potency


Core dependence: release & size

0

20

40

60

80

100

0 240 480 720 960

Time(min)

% R

elea

sed

Drug release in 0.1N HCl, same sized cores and same

coat thickness

Starting PSDs and ghost PSDs after drug release

Heinicke & Schwartz, Pharm Dev Tech, 12, 473, 2007


Future opportunities for DIA

Count, mass per particle, uniformitySurrogate dissolution testSurface area of core chargeBimodal distributions Density of drug layer or coating layer as appliedApplication to other multi-particulates


Acknowledgements

Horiba instrumentsThe late Dr. J.B. Schwartz & USPDr. Garth Boehm and Dr. Kristin ArnoldAnchen Pharmaceuticals Inc.

Questions?

Dynamic Image Analysis and pharmaceutical multi-particulates · Dynamic Image Analysis and...

Documents

Transcript of Dynamic Image Analysis and pharmaceutical multi-particulates · Dynamic Image Analysis and...