who guidelines onwho guidelines on stability evaluation · pdf filewho guidelines onwho...

WHO Guidelines onWHO Guidelines on Stability Evaluation of Vaccines

Presented by Tim SchofieldDirector, US Regulatory AffairsGSK BiologicalsGSK Biologicals

WHO Guidelines on Stability Evaluation of Vaccines

The views and opinions expressed in the following PowerPoint slides are those of the individual presenter and should not be attributed to Drug Information Association, Inc. (“DIA”), its directors, officers, employees, volunteers, members, chapters, councils, Special Interest Areavolunteers, members, chapters, councils, Special Interest Area Communities or affiliates, or any organization with which the presenter is employed or affiliated.

These PowerPoint slides are the intellectual property of the individual presenter and are protected under the copyright laws of the United States of America and other countries. Used by permission. All rights reserved. Drug Information Association DIA and DIA logo are registered trademarks orInformation Association, DIA and DIA logo are registered trademarks or trademarks of Drug Information Association Inc. All other trademarks are the property of their respective owners.

2www.diahome.orgDrug Information Association

Outline

• Overview of WHO GuidelineOverview of WHO Guideline• Stability quality attributes• Basic principles of vaccine stabilityBasic principles of vaccine stability• Stability during development• Stability supporting licensure• Stability supporting licensure• Post licensure stability evaluation• Challenges to implementation• Challenges to implementation

Drug Information Association www.diahome.org 3

Overview of WHO Guidelines

• Acknowledges the importance of stability to the success of immunizationstability to the success of immunization programs worldwide

• Provides a scientific basis and guiding g gprinciples for evaluation of stability over the vaccine lifecycle– For the purpose of clinical trial

monitoringg– For licensing– For post licensure monitoring

• Adopted by the 57th meeting of the• Adopted by the 57th meeting of the WHO Expert Committee on Biological Standarization, 23-27 October 2006


Overview of WHO Guidelines (cont.)

• Supported by pp yimplementation workshops

Seoul Korea (Apr 2008)– Seoul, Korea (Apr 2008)– Geneva, Switz. (Oct 2008)

• Workshop proceeding published in a special issue of Biologicalsissue of Biologicals, November 2009, 37(6)


Stability Quality Attributes

• Stability quality attributes should include those properties which impact safety and/or efficacywhich impact safety and/or efficacy– e.g., potency, sterility, etc.

• Note: all properties change over time; thus any parameter related to safety and/or efficacy should be part of the vaccine stability program

• Stability quality attributes should also include propertiesStability quality attributes should also include properties which impact stability over the course of shelf-life– e.g., increase in moisture over time for a lyophilized vaccine

• Similarly properties which impact stability should be part of the release specification for the product– Moisture of a lyophilized vaccineMoisture of a lyophilized vaccine– pH of an adjuvanted vaccine


Basic Principles of Vaccine Stability

• A scientific basis of stability begins with d t di h i d dunderstanding how vaccines degrade

– First order kinetics• The rate of decay is [C] dependent Po

tenc

y

td d tik

potency, initialP where

,ePPotency

0

tk0

=

⋅= ⋅−

0 6 12 18 24 30 36

Time (Months)

• Linear in log potency

rate.ndegradatiok =

tk)Pl ()P tl ( (Pot

ency

)

– The log transformation also “normalizes” potency measurements, and “stabilizes”

i bilit th t

tk)Pln()Potencyln( 0 ⋅−=

0 6 12 18 24 30 36

Time (Months)

ln(

variability across the potency range


Basic Principles of Vaccine Stability (cont.)

• A first order kinetics equation (log of potency) is fit to vaccine stability data using least squares regression 3 2

3.43.63.84.0

/mL

Me Stability

3 23.43.63.84.0

/mL

Me Stability

q g• Like all statistical estimates,

the least squares regression ti i i t d ith

2.02.22.42.62.83.03.2

log 1

0TC

ID50

/

2.02.22.42.62.83.03.2

log 1

0TC

ID50

/

equation is associated with variability– This can be expressed as a

0 6 12 18 24 30 36Time (Months)

0 6 12 18 24 30 36Time (Months)

confidence interval on the regression line

– Forms the basis for ICH shelf-life determinationdetermination


Basic Principles of Vaccine Stability (cont.)

• Study design should acknowledge the goal of the stability studystability study– ICH intervals are designed to provide sufficient data at time of filing– Statistical design can be used to minimize uncertainty

4.95.15.35.5

mL

Usual ICH Intervals– Shelf-life Determination

• Data clustered at the desired shelf-life will minimize impact of uncertainty on SL

– Determination of loss rate• Testing at beginning and end will

reduce uncertainty on the loss rate

3.43.63.84.0

mL

Me Stability

3.53.73.94.14.34.54.7

0 6 12 18 24 30 36

log 1

0TC

ID50

/m

5.5

Centered on 24-Months

uncertainty on SL determination

Design 1

Design 2

2.02.22.42.62.83.03.2

0 6 12 18 24 30 36

log

10T

CID

50/m

0 6 12 18 24 30 36Time (Months)

3.94.14.34.54.74.95.15.3

log 1

0TC

ID50

/mL

0 1 2 3 4 5

Design 1 Design 21 12 13 44 4

0 6 12 18 24 30 36Time (Months)


3.53.73.9

0 6 12 18 24 30 36Time (Months)

4 4sb=0.45 sb=0.33

Reduction=25%

Stability During Development

• Strategic use of accelerated stability dataU d t di i t bilit– Understanding vaccine stability

• Mechanism of degradation• Kinetics model

– Formulation development– Impact of bulk stability on final product stability

• in lieu of sequential stability

– Benchmark for vaccine changes 0

1

Arrehnius Plot

changes• Process change• Facility change

-8-7-6-5-4-3-2-10

ln(k

)


-9-8

0.003 0.0032 0.0034 0.00361/|T|

T 60C 45C 30C 22C 10C 5C

Stability During Development (cont.)

• Use clinical stability to define what the subject

Clinical Stability

Regression

Upper Bound

received, and thereby specifications– Using immunogenicity

Pote

ncy

as the endpoint, interpolate the potency associated with a clinical correlate of

Res

pons

e

0 6 12 18 24Time (Months)

clinical correlate of efficacy

– Using efficacy as the endpoint, perform a

Dose

80%

100%

p , plogistic analysis and interpolate the dose corresponding to

d i d ffi 0%

20%

40%

60%Protected

Unprotected

Logistic

a desired efficacy claim


0%

Dose

Stability Supporting Licensure

• Shelf-life determination –Long Term Stability at 2-8 C

4 0

measles example– Is shelf-life 12-months due

to a stability measurement EAC

2 5

3.0

3.5

4.0

y (lo

g10

TCID

50/D

ose)

Lot 1

Lot 2

Lot 3EACto a stability measurement at 18-months for lot 2 which falls below the expiry acceptance criteria (EAC)?

• “Compliance Model”

2.0

2.5

0 6 12 18 24 30 36

Time (Months)

Pote

ncy

Shelf-Life = 12-months

• Compliance Model– ICH Q1E defines shelf-life

as the time where the lower bound on the confidence

Shelf-Life Determination (All Lots)

3.6

3.8

4.0

mL

Lot 1Lot 2Lot 3RegressionLower BoundMin. Req.

interval intersects the EAC• “Estimation Model”• Risk based approach

2 6

2.8

3.0

3.2

3.4lo

g 10T

CID 5

0/

Shelf-Life = 24-months

EAC


2.60 6 12 18 24 30 36

Time (Month)

Stability Supporting Licensure (cont.)

• Note: Shelf-life determination does not acco nt for ariabilit indoes not account for variability in release potencies of future manufactured lots

A manufactured lot released

Rel

ease

Lot 1

Development

– A manufactured lot released below the stability lots will have EAC before end of shelf-life Manuf

0m0m 24m

Lot 2Lot 3

EAC

<24m

• A minimum release limit assures EAC by end of shelf-life– Calculated from combination of

LabelingPackaging

Shipping (15 C)

Labeled Storage (2-8 C) Use – R/S2-8°C– Calculated from combination of

accumulated losses over shelf-life, together with statistical uncertainties

g gInspection

(22 C)

( )

∑ ∑ +++= 2Assay

2b

2idfii ssttbtEACleaseRe Minimum

i


Post Licensure Stability Evaluation

• Similar to shelf-life determination, stability modeling should be tili d t ti t d t lit d i t bilit it iutilized to estimate product quality during stability monitoring– Highly variable measurements yield sporadic stability OOS results– The stability model yields a more precise estimate of vaccine quality

Regression versus Individual Stability Time PointRegression versus Individual Stability Time Point

ten

cyte

ncy

PoPo

t


0 6 12 18 24 30Time (Months)

0 6 12 18 24 30Time (Months)

Post Licensure Stability Evaluation (cont.)

• Stability comparison after a process or facility change– Stability is a product quality attribute Dist ib tion ofStability is a product quality attribute

– Distribution modeling can be used to determine an acceptable change in

Distribution ofloss rates

2-8ºC2-8ºC

Rel

ease

Distribution ofRelease potencies

p gstability rate

– Using distributions of release, slope, and time to administration 15

20

Distribution oftime to administration

0m0m 24m

Labeled Storage

– Can determine the distribution of expiry potencies; a shift in the distribution of expiry potency can be used to derive a limit on the

0

5

10

ΔΔPotencybe used to derive a limit on the change in degradation rate

– Accelerated stability can be used to facilitate an early evaluation of a

Parallel Arrhenius Plots for New and Old Process Materials

-2-10123

ln(S

lope

) Old

New


change in stability -4-3

3.1 3.2 3.3 3.4 3.5 3.6 3.7

1000/(Co+273|

(45o) |

(37o) |

(22o) |

(15o) |

(5o) |

Challenges to Implementation

• Statistical thinking and modeling– Appreciation of variability and risk– Growing awareness of the need for skilled

statisticians in nonclinical developmentstatisticians in nonclinical development• Statistical approaches to bioassay development,

validation, and maintenance• Application of design of experiments to support quality

by design• Statistical process controlStatistical process control• Stability modeling and comparability strategies

– Statistical training of industry and regulatory scientists– User friendly software solutions


Challenges to Implementation (cont.)

• Inaccurate stability modeling can lead to poor estimates of vaccine shelf-lifeof vaccine shelf life– The default model for

stability of vaccines is a 1st order kinetics model

Stability Kinetics0‐order

1st order

1 order kinetics model– Modeling by 0-order

kinetics can lead to underestimation of shelf- 0 6 12 18 24underestimation of shelflife, and limitations on vaccine supply

– Some vaccines degrade 4.505.005.50

cyt0031.0t0840.0

tb2

tb1

e72.1e43.3

eaeaPotency log 21

⋅−⋅−

⋅⋅

⋅+⋅=

⋅+⋅=

Time (Month)

Some vaccines degrade by higher order kinetics, leading to complex stability modeling 2.00

2.503.003.504.00

0 6 12 18 24

ln P

oten

c


y g0 6 12 18 24

Time (Months)


• Harmonization of stability modeling and stability monitoringICH shelf life determination ses a model of the mean prod ct

Observations

Regression

X Observations

Regression

X ObservationsRegression

X– . . . however, stability OOS lt it d d i t

– ICH shelf-life determination uses a model of the mean product stability profile

g

Lower 95% CI

X

X

X

X

X

Regression

Lower 95% PI

X

X

X

X

X

Lower 95% CILower 95% PI

X

X

X

X

X

results are cited during post licensure studies• Ex., a batch which yields a

24 month shelf life pre

0 4 8 12 16 20 24 28 32

X

X

0 4 8 12 16 20 24 28 32

X

X

0 4 8 12 16 20 24 28 32

5%9%

18%

X

X24-month shelf-life pre-licensure would have a ~30% chance of yielding a stability OOS if tested post

Time (Months)Time (Months)0 4 8 12 16 20 24 28 32

Time (Months)

Probability of OOS= 1-(0.95x0.91x0.82) = 29%

licensure– Post licensure data should

be statistically modeled to

Probability of OOS=1 – (0.95*0.91*0.82) = 29%


reduce risk of failing a good lot


• Application to legacy products which are controlled t t tto target– Legacy vaccine specifications are typically established to

assure consistency at releaseassure consistency at release• No provision for product stability• Release and EAC are the same

Th WHO G id li– The WHO Guidelines should be applied to vaccines which have SpecificationsRelease

LimitsCapability

Limits

been developed with a vision towards supporting release and

Shelf-LifeMinimum Requirement


supporting release and expiry requirements

-6 0 6 12 18 24 30 36

Summary

• The WHO Guidelines on Stability Evaluation of Vaccines yprovides a scientific framework for assuring vaccine quality throughout shelf-lifeA i t t ti ti l d i d l i d th• Appropriate statistical design and analysis reduces the uncertainty in vaccine stability evaluation, and thereby risk

• Implementation of the guidelines has both statistical and practical challenges which must be addressed to help assure adequate supply of quality vaccines to the worldassure adequate supply of quality vaccines to the world


who guidelines onwho guidelines on stability evaluation · pdf filewho guidelines onwho...

Documents