Documente Academic
Documente Profesional
Documente Cultură
ICH Guideline Q1A defines Shelf Life (also referred to as ICH Guideline Q1E defines shelf life as
expiration dating period) as
The shelf life of a pharmaceutical product is the maximum
The shelf life for a pharmaceutical product is the maximum time at which the true mean response of a stability limiting
time at which a stability limiting characteristic stays within characteristic crosses the acceptance criterion.
acceptance
p criteria.
basis for the current ICH/FDA shelf life estimation procedure
Also in Q1A, Specification Shelf Life is defined as limited assurance that individual test results will comply with
the specification up to m months
The combination of physical, chemical biological, and
microbiological tests and acceptance criteria that determine focus on the mean response implies the risk to fail specification
the suitability of a drug substance throughout its re-test at shelf life will be 50%
period, or that a drug product should meet throughout its
shelf life.
Pharmaceutical Stability Shelf Life August 1, 2010 2 Pharmaceutical Stability Shelf Life August 1, 2010 3
ICH Shelf Life Estimation No Pooling ICH Shelf Life Estimation Pooled Batches
Confidence Band
Response
Response
Stability Limiting R
Pharmaceutical Stability Shelf Life August 1, 2010 8 Pharmaceutical Stability Shelf Life August 1, 2010 9
usually a simple linear (straight line) regression model is ICH methodology suggest to test for equal regression slopes
assumed to characterize the response-time continuum among stability batches first
first-order nonlinear models are more appropriate for some if batch slopes are nonsignificant ( = 0.25)
stability limiting characteristics common regression slope is assumed among batches
we will focus on simple linear regression models p are tested
batch intercepts
simple linear model if batch intercepts are nonsignificant ( = 0.25)
yij = b0i + b1i (monthj) + ij - common intercept is assumed among batches
- batches are pooled
yij = observation at jth month for ith batch if batch slopes are significant, no further testing is considered
b0i = batch intercept and b1i = batch regression slope ICH does not allow for a model with a common batch intercept
ij = residual error with Normal assumptions and unequal slopes (which can be an important model)
Pharmaceutical Stability Shelf Life August 1, 2010 10 Pharmaceutical Stability Shelf Life August 1, 2010 11
ICH Estimation Issues ICH Estimation Issues
115
26 stability batches
all on same product 110
105
90
85
0 6 12 18 24 30 36 42 48
Storage Time (Months) Product Shelf Life
Pharmaceutical Stability Shelf Life August 1, 2010 14 Pharmaceutical Stability Shelf Life August 1, 2010 15
Comparison of ICH Shelf Life Estimation Methodology Using Industry Data Comparison of ICH Shelf Life Estimation Methodology Using Industry Data
Using 3 Batches with 24 Months of Data, Extrapolating to 48 Months of Storage Time Using 6 Batches with 24 Months of Data, Extrapolating to 48 Months of Storage Time
Considering All Possible Response Models Considering All Possible Response Models
Product Shelf Life = 37.8 Product Shelf Life = 37.8
450 3500
16.31%
15.27% 15.54%15.34%
400 15.03%
14.42% 3000
350 13.17%
2500 11.94%
300
Frequency
Frequency
250 2000
8.92% 8.58%
7.81%
200 7.08% 1500 6.83%
6.35% 6.13%
150 5.50%
4.96%
1000
3.85%
100 3.22%
2.46% 2.31%
1.88% 500 2.00%
50 1.27% 1.43%
0.92%
0.08% 0.31% 0.46% 0.35% 0.20% 0.08% 0.04%
0.00%
0 0
12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0 12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0
ICH Estimated Shelf Life (Total = 2600) ICH Estimated Shelf Life (Total = 230230)
Pharmaceutical Stability Shelf Life August 1, 2010 18 Pharmaceutical Stability Shelf Life August 1, 2010 19
141
Pharmaceutical Stability Shelf Life August 1, 2010 20 Pharmaceutical Stability Shelf Life August 1, 2010 21
2500 250
2258
2000 200
1500 150
Frequency
Frequency
113
1000 100 109
97
870 87
80 82
704
500 566 50 54
45 43
35
216 243 25
129
0 0
12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0 12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0
ICH Estimated Shelf Life (Total = 230230) ICH Estimated Shelf Life (Total = 2600)
Pharmaceutical Stability Shelf Life August 1, 2010 22 Pharmaceutical Stability Shelf Life August 1, 2010 23
6-Batch Estimate of Shelf Life 3-Batch Estimate of Shelf Life
n = 6015 (30%) mean = 20.1 months SD = 4.13 n = 983 (38%) mean = 27.8 months SD = 4.61
Comparison of ICH Shelf Life Estimation Methodology Using Industry Data Comparison of ICH Shelf Life Estimation Methodology Using Industry Data
Using 6 Batches with 24 Months of Data, Extrapolating to 48 Months of Storage Time Using 3 Batches with 24 Months of Data, Extrapolating to 48 Months of Storage Time
Considering All Possible Response Models, Model_01: Common Intercept Considering All Possible Response Models, Model_10: Common Slopes
Product Shelf Life = 37.8 Product Shelf Life = 37.8
2500 250
252
186
164
1500 150
Frequency
Frequency
1000 100
1009
94
739 765 72
683 68
500 50
48
324 34
218
17
109 11
0 0
12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0 12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0
ICH Estimated Shelf Life (Total = 230230) ICH Estimated Shelf Life (Total = 2600)
Pharmaceutical Stability Shelf Life August 1, 2010 24 Pharmaceutical Stability Shelf Life August 1, 2010 25
2500 250
2413
2274
2000 200
1500 150
Frequency
Frequency
1177
1000 1080 100
63
500 554 50
49
44 43 43
235 269 28 26
20 18
159 120 13 11
0 0
12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0 12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0
ICH Estimated Shelf Life (Total = 230230) ICH Estimated Shelf Life (Total = 2600)
Pharmaceutical Stability Shelf Life August 1, 2010 26 Pharmaceutical Stability Shelf Life August 1, 2010 27
Poolability
Using 6 Batches with 24 Months of Data, Extrapolating to 48 Months of Storage Time
Considering All Possible Response Models, Model_00: Common Intercepts and Slopes
Product Shelf Life = 37.8
1000
use = 0.25 level of significance
estimate of shelf life too heavily dependent on best model
500 assuming unequal slopes forces the shelf life estimate to be
based on worst batch
125
0
12.0 14.0 16.0 18.0 20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0 tends to minimize shelf life estimate
ICH Estimated Shelf Life (Total = 230230)
Pharmaceutical Stability Shelf Life August 1, 2010 28 Pharmaceutical Stability Shelf Life August 1, 2010 29
Random Batch Mixed Model Analysis Random Batch Mixed Models - Reflection
Response
mixed model analysis would Band
Stability Limiting R
model between-batch variation as a random effect Mean of Random Batch
quantify both between and within-batch variation separately Mixed Model Analysis
Pharmaceutical Stability Shelf Life August 1, 2010 30 Pharmaceutical Stability Shelf Life August 1, 2010 31
3-Batch Estimate of Shelf Life Confidence Interval 6-Batch Estimate of Shelf Life Confidence Interval
Alternatives to ICH Shelf Life Estimation Methodology Using Industry Data Alternatives to ICH Shelf Life Estimation Methodology Using Industry Data
Using 3 Batches with 24 Months of Data, Extrapolating to 48 Months of Storage Time Using 6 Batches with 24 Months of Data, Extrapolating to 48 Months of Storage Time
Estimating Shelf Life Using Random Batch Mixed Model Analysis Estimating Shelf Life Using Random Batch Mixed Model Analysis
Product Shelf Life = 37.8 Product Shelf Life = 37.8
450 3500
16.47%
15.76%
400 15.00% 15.01%
3000
13.59%
350 12.65%
11.65% 2500
11.46% 11.37%
300
Frequency
Frequency
200 1500
6.65%
150
4.62% 1000 4.69%
4.31%
100
2.54% 2.23% 500 2.12%
50 1.31% 1.12% 1.38%
1.00% 0.81%
0.38%
0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00%
0 0
12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44
Reflection Calibration Method (Total = 2600) Reflection Calibration Method (Total = 230230)
Pharmaceutical Stability Shelf Life August 1, 2010 32 Pharmaceutical Stability Shelf Life August 1, 2010 33
Random Batch Mixed Models Distribution of -hat 3-Batch Estimate of Shelf Life Confidence Interval
400
14.46%
13.69%
Response
350
12.46%
300
Stability Limiting R
9.88%
Frequency
100 3.50%
2.42% 2.46%
1.96%
1.54%
Labeled Calibration 50
Pharmaceutical Stability Shelf Life August 1, 2010 34 Pharmaceutical Stability Shelf Life August 1, 2010 35
6-Batch Estimate of Shelf Life Confidence Interval Conclusions
several advantages
3500
16.32%16.27%
can be used to extend the inference of a shelf life statement to
3000
13.72% future batches
depending on how well the stability data represent the
12.54%
2500 12.16%
11.71%
manufacturing process
Frequency
2000
1500 7.19%
avoids the batch poolability issue
breaks the dependence of the estimate of shelf life on which
reduced or pooled regression model is selected
1000 4.23%
3.49%
500
1.44%
0.88%
avoids the issue of a common intercept / different slope
0
0.00% 0.00% 0.00% 0.00% 0.00% 0.09%
regression model being appropriate or not
12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44
B_Hat Calibration Method (Total = 230230)
Pharmaceutical Stability Shelf Life August 1, 2010 36 Pharmaceutical Stability Shelf Life August 1, 2010 37
Conclusions Conclusions
Pharmaceutical Stability Shelf Life August 1, 2010 38 Pharmaceutical Stability Shelf Life August 1, 2010 39