Determination of Water Vapor Transmission Rate for Various High Barrier

Blister Packs
Executive Summary

Water Vapor Transmission Rate (WVTR) testing was performed to construct weight gain
profiles due to moisture ingress of formed and sealed blisters (PVC, PVDC, Aclar Ultrx 2000
and cold form foil) when stored at 23°C/75% and 40°C/75% RH. The purpose of the test was to
determine the potential to improve USP <671> by determining the time needed to reach steady
state, the duration, and number of data points needed for WVTR calculation, the variability in
data due to samples and testing labs, and to determine if empty blisters (i.e. “controls”) are
needed for accurate WVTR determination. This report summarizes the data collected from four
testing labs - Merck, Pfizer, sanofi-aventis and Abbott Labs. Mean WVTR results determined by
linear regression without zero time point is given in the following table:

Table 1
WVTR Means (mg/cavity/day)

Condition Site PVC PVdC Ultrx Cold

23C/75%RH AB 1.215 0.185 0.029 0.005

MK . 0.138 0.028 0.003
PF 1.223 0.127 0.028 -0.001
SA 1.280 0.142 0.028 -0.004

40C/75%RH AB 2.978 0.789 0.136 0.011

MK . 0.791 0.137 0.003
PF 3.183 0.882 0.152 0.008
SA 3.019 0.841 0.144 -0.007

The primary findings are as follows:

 Steady state appears to be established after the second time point (first time point after
initial) with the exception of PVC.
 The recommended time duration for testing is 2 days for PVC and 35 days for barrier
blisters.1.
 The recommended number of test points is 2 for PVC and 5 time points for barrier
blisters.
 The controls have little effect on the WVTR and can be safely eliminated.
 Some blister types and the controls show a higher increase in moisture from the first time
point (at time ‘0’) to the second time point than over the remaining time points.
 There is noticeable curvature in the PVC weights over time.
 There is no change over time in cold form blisters.
 For each blister type, site slopes are similar.

1
Throughout this paper, blisters formed from PVdC, Aclar, and cold form foil are referred to as barrier blisters, as
distinguished from PVC blisters.

1
  Both conditions of 23C/75%RH and 40C/75%RH discriminate between the blister types
with better discrimination at the higher temperature.

Based on the results of this study, we recommend:

 The use of 40ºC/ 75% RH storage condition based on reduced variability as measured by
RSD.
 The use of the ASTM E 96 calculator to ensure a constant partial moisture vapor pressure
differential, determine the required number of blisters and balance sensitivity to ensure
data precision and accuracy. Refer to Appendix 1.
 A test duration of 2 days for PVC and 35 days for barrier blisters at 40ºC/ 75% RH.
 The use of 2 time points for PVC at initial and day 2 and the use of 5 (weekly) time
points for barrier blisters beginning at day 7. Eliminate the initial time point (time ‘0’)
for barrier blisters.
 Eliminate the use of empty controls when using non-paper backed lidding foil.
 The use of linear regression for calculating WVTR for barrier blisters.
 Report results in mg/day/blister cavity and eliminate reference to blister classification to
enable comparison with other container closure systems (e.g. bottles).

Introduction & Objectives

The purpose of this study was to assess the opportunity to improve USP <671> by constructing
the weight gain profiles, due to moisture ingress, of formed and sealed blisters when stored at
23°C/75% and 40°C/75% RH. The profiles were used to determine the time needed to reach
steady state, the duration, and number of data points needed for Water Vapor Transmission Rate
(WVTR) calculation, the variability in data due to samples and testing labs, and to determine if
the empty blisters (i.e. the “control”) are needed for accurate WVTR determination. The origins
of this study also lie in part in a previous publication1 that provided a theoretical basis for using
WVTR per unit dose as a means to compare different container-closure systems.

Materials & Methods

Blisters were formed and filled with an appropriate amount of desiccant (or without desiccant in
the case of controls) determined using the ASTM E 96 calculator described in Appendix 1, and
sealed to a heat sealable foil lid structure according to established procedures. Subsequently, the
blister samples were stored in environmental chambers controlled at 40±2°C/75±5%RH or
23±2°C/75±3%RH. Each individual container (blister pack or set of blister packs) was weighed
at various time points throughout the study according to a study protocol as summarized in
Appendix 2. The weights of the blisters were tabulated for statistical analysis and for
determination of WVTR (in mg per day per blister cavity) for the specific blister types (material
& cavity design) at the two testing conditions. A linear line of best fit was applied to the weight
data for each container using linear regression. The slope of each line is the WVTR.

2
Four blister types were used in the study: (1) 200 µ PVC, (2) 200 µ PVC/ 60 gsm PVdC, (3) 200
µ PVC/ 51 µ Aclar, and (4) cold form foil. The lidding foil was 20 µ Push-Thru, without paper
or supporting film. Details on test duration, weighing intervals, amount of desiccant, number of
cards, number or replicates, etc are given in Appendix 2. The blisters were formed on a Pentapak
CT1200 using toolings designed to accommodate a size ‘0’ capsule. Four labs performed the
WVTR testing: Merck (MK), Pfizer (Pf), sanofi-aventis (SA), and Abbott (AB).

Results/Discussion
A number of containers (n=17 of a total 300 blister cards) were removed from the analysis due to
holes found in the blisters. All of these containers contained desiccant. Merck did not perform
the PVC testing. Storage conditions were reversed for the Abbott blisters during days 15-17.

As can be seen in Appendix 2, the number of blister packs was not consistent across the four
blister types. To compare across blister types, results were converted to mg/cavity. Note that
for PVC, AB performed two runs of 5 containers apiece denoted AB1 & AB2 in the figures.

Summary

The weight and the calculated WVTR for each blister type are given in the following subsections
(A-D). Each subsection consists of two figures followed by a slope summary table.

Figure descriptions:

Figure #1: Mean weight (mg/cavity) of 10 containers with desiccant at each time point by
lab. Blisters were stored at 23C/75%RH.

Figure #2: Mean weight (mg/cavity) of 10 containers with desiccant at each time point by
lab. Blisters were stored at 40C/75%RH.

Slope Summary Table Description

For each container type, there are two sets of results within the slope summary table. Each set
provides the number of slopes (containers) used in the calculation, the mean, standard deviation,
and RSD(%) of the slopes as well as the minimum and maximum slope for each site and
condition. The two sets are: 1) Containers with desiccant and 2) Containers with desiccant after
adjusting for the mean of the empty controls (i.e., At each time point, the mean of the empty
controls (usually n=10) was subtracted from the weight of each container containing desiccant
prior to fitting the linear regression line to determine the WVTR).

3
A) PVC
Figure A1 Figure A2
23C/75% RH 40C/75%RH

Table A
PVC Blisters
Slope Summary Statistics

Set 1 - As Is Mean Std Dev Minimum Maximum

Condition Site N (mg/cavity/day) (mg/cavity/day) RSD(%) Slope Slope
23C/75%RH AB 10 1.2230 0.0103 0.8402 1.2086 1.2371
23C/75%RH PF 10 1.2313 0.0167 1.3552 1.2107 1.2690
23C/75%RH SA 10 1.2943 0.0088 0.6803 1.2840 1.3074
40C/75%RH AB 10 3.0820 0.0973 3.1586 2.8903 3.2236
40C/75%RH PF 10 3.2391 0.0387 1.1957 3.1858 3.2951
40C/75%RH SA 9 3.1387 0.0919 2.9282 2.9230 3.2429

Set 2 (Adjusted for Controls) Mean Std Dev Minimum Maximum

Condition Site N (mg/cavity/day) (mg/cavity/day) RSD(%) Slope Slope
23C/75%RH AB 10 1.2241 0.0101 0.8279 1.2090 1.2375
23C/75%RH PF 10 1.2256 0.0167 1.3615 1.2049 1.2633
23C/75%RH SA 10 1.2901 0.0088 0.6826 1.2798 1.3032
40C/75%RH AB 10 3.0807 0.0966 3.1368 2.8899 3.2213
40C/75%RH PF 10 3.2297 0.0387 1.1992 3.1763 3.2856
40C/75%RH SA 9 3.1330 0.0919 2.9335 2.9174 3.2372

4
B) PVdC
Figure B1 Figure B2
23C/75% RH 40C/75%RH

Table B
PVdC Blisters
Slope Summary Statistics

Set 1 - As Is Mean Std Dev Minimum Maximum

Condition Site N (mg/cavity/day) (mg/cavity/day) RSD(%) Slope Slope
23C/75%RH AB 10 0.2067 0.0307 14.855 0.1869 0.2671
23C/75%RH MK 10 0.1448 0.0010 0.674 0.1425 0.1458
23C/75%RH PF 10 0.1366 0.0078 5.736 0.1150 0.1407
23C/75%RH SA 10 0.1464 0.0006 0.435 0.1453 0.1472
40C/75%RH AB 10 0.8207 0.0137 1.663 0.7887 0.8366
40C/75%RH MK 10 0.8277 0.0161 1.948 0.7906 0.8475
40C/75%RH PF 10 0.9146 0.0102 1.117 0.8953 0.9284
40C/75%RH SA 10 0.8764 0.0091 1.034 0.8592 0.8901

Set 2 - Adjusted for Controls Mean Std Dev Minimum Maximum

Condition Site N (mg/cavity/day) (mg/cavity/day) RSD(%) Slope Slope
23C/75%RH AB 10 0.1929 0.0307 15.9124 0.1731 0.2534
23C/75%RH MK 10 0.1316 0.0010 0.7420 0.1293 0.1325
23C/75%RH PF 10 0.1288 0.0078 6.0823 0.1072 0.1329
23C/75%RH SA 10 0.1347 0.0006 0.4731 0.1337 0.1356
40C/75%RH AB 10 0.8096 0.0137 1.6862 0.7775 0.8255
40C/75%RH MK 10 0.8158 0.0161 1.9765 0.7787 0.8357
40C/75%RH PF 10 0.9031 0.0102 1.1317 0.8838 0.9169
40C/75%RH SA 10 0.8667 0.0091 1.0455 0.8495 0.8804
5
C) Ultrx 2000
Figure C1 Figure C2
23C/75% RH 40C/75%RH

Table C
Ultrx 2000 Blisters
Slope Summary Statistics
Set 1 - As Is Mean Std Dev Minimum Maximum
Condition Site N (mg/cavity/day) (mg/cavity/day) RSD(%) Slope Slope
23C/75%RH AB 10 0.0306 0.0104 33.891 0.0240 0.0562
23C/75%RH MK 10 0.0289 0.0038 13.027 0.0262 0.0366
23C/75%RH PF 10 0.0287 0.0064 22.428 0.0252 0.0452
23C/75%RH SA 7 0.0283 0.0008 2.830 0.0276 0.0296
40C/75%RH AB 10 0.1389 0.0020 1.465 0.1355 0.1423
40C/75%RH MK 7 0.1387 0.0012 0.857 0.1377 0.1407
40C/75%RH PF 10 0.1563 0.0204 13.080 0.1396 0.1928
40C/75%RH SA 10 0.1463 0.0017 1.185 0.1430 0.1479
Set 2 - Adjusted for Controls Mean Std Dev Minimum Maximum
Condition Site N (mg/cavity/day) (mg/cavity/day) RSD(%) Slope Slope
23C/75%RH AB 10 0.0289 0.0104 35.9286 0.0222 0.0544
23C/75%RH MK 10 0.0264 0.0038 14.2965 0.0237 0.0341
23C/75%RH PF 10 0.0267 0.0064 24.0997 0.0232 0.0432
23C/75%RH SA 7 0.0258 0.0008 3.1073 0.0251 0.0271
40C/75%RH AB 10 0.1363 0.0020 1.4935 0.1329 0.1397
40C/75%RH MK 7 0.1360 0.0012 0.8746 0.1350 0.1379
40C/75%RH PF 10 0.1543 0.0204 13.2463 0.1376 0.1909
40C/75%RH SA 10 0.1437 0.0017 1.2061 0.1404 0.1453

6
D) Cold Form
Figure D1 Figure D2
23C/75% RH 40C/75%RH

Table D
Cold Form Blisters
Slope Summary Statistics
Set 1 - As Is Mean Std Dev Minimum Maximum
Condition Site N (mg/cavity/day) (mg/cavity/day) RSD(%) Slope Slope
23C/75%RH AB 9 0.0069 0.0160 231 -0.0030 0.0443
23C/75%RH MK 8 0.0031 0.0007 23 0.0023 0.0044
23C/75%RH PF 10 0.0051 0.0154 301 -0.0040 0.0364
23C/75%RH SA 8 -0.0006 0.0111 -1896 -0.0057 0.0265
40C/75%RH AB 10 0.0110 0.0240 218 -0.0064 0.0575
40C/75%RH MK 6 0.0020 0.0058 296 -0.0012 0.0134
40C/75%RH PF 10 0.0146 0.0218 149 -0.0085 0.0423
40C/75%RH SA 9 -0.0069 0.0032 -46 -0.0089 -0.0008
Set 2 - Adjusted for Controls Mean Std Dev Minimum Maximum
Condition Site N (mg/cavity/day) (mg/cavity/day) RSD(%) Slope Slope
23C/75%RH AB 9 0.0081 0.0160 198 -0.0019 0.0455
23C/75%RH MK 8 0.0002 0.0007 433 -0.0007 0.0014
23C/75%RH PF 10 0.0093 0.0154 166 0.0002 0.0406
23C/75%RH SA 8 0.0051 0.0111 219 -0.0001 0.0322
40C/75%RH AB 10 0.0173 0.0240 138 -0.0001 0.0638
40C/75%RH MK 6 0.0040 0.0058 144 0.0009 0.0154
40C/75%RH PF 10 0.0237 0.0218 92 0.0006 0.0514
40C/75%RH SA 9 0.0017 0.0032 185 -0.0002 0.0079
7
The following table summarizes the observations from the figures and tables.

Table 2 Findings Summary for Blister Types and Test Sites

Blister Type
PVC PVDC Ultrx 200 Cold Form
Curvature 40C/75%RH - Yes 40C/75%RH - Some No No
over time? 23C/75%RH - No 23C/75%RH - No
Adjusting Little effect (< 0.01 Little effect. (< 0.014 Reduces slope No effect
for Controls mg/cavity/day) mg/cavity/day) slightly. (<0.003
mg/cavity/day)
Does not remove
curvature.
Control Slight increase from Slight increase from Slight increase No increase over
Profile initial to first time initial to first time from initial to first time.
point after initial at point after initial at time point after
40C - then no 40C - then no initial - then no
increase to final increase to final time increase to final
time point. point. time point.

PF has higher MK controls have

intercept. higher intercept
than other sites.
Intercepts SA has higher Site intercepts Similar at 40C. Similar intercepts at
(Blisters intercept at 40C. similar. Slight differences 23C. Some
with at 23C/75. variation at 40C.
desiccant)
Slopes Site slopes similar Site slopes similar to Site slopes similar All slopes close to
(Blisters to one another. one another. to one another. zero.
with
desiccant) 40C slopes higher 40C slopes higher 40C slopes higher
than 23C slopes. than 23C slopes. than 23C slopes.

Within run standard Standard deviations Standard

deviations around about regression lines deviations about
regression line at 23C vary from regression lines at
similar. Possible run container to container 40C vary from
to run slope and for PF. container to
standard deviation container for PF.
around regression
line variation based
on Abbot data.

As noted in the overview, the formed blisters were sealed with 20 µ Push-Thru lidding foil without paper or
supporting film. It is thought that avoiding a hygroscopic paper layer contributed to the negligible difference
between WVTR calculated with and without the use of controls. Therefore use of non-paper backed foil is
regarded as a necessary condition to eliminate the use of controls.

8
Comparison of Blisters and Labs

Table 3 and Figures 1 & 2 show the average slope for each blister type by site combination. Notice that site
results are similar for each blister type with the blister type slopes decreasing as the blister barrier increases.
Table 4 shows the standard deviation of the slopes for each combination.

Table 3. Comparison of WVTR for Blister Types and Test Sites (Mean, mg/cavity/day)

Condition Site PVC PVdC Ultrx Cold

23C/75%RH AB 1.2230 0.2067 0.0306 0.0069

23C/75%RH MK . 0.1448 0.0289 0.0031
23C/75%RH PF 1.2313 0.1366 0.0287 0.0051
23C/75%RH SA 1.2943 0.1464 0.0283 -0.0006

40C/75%RH AB 3.0820 0.8207 0.1389 0.0110

40C/75%RH MK . 0.8277 0.1387 0.0020
40C/75%RH PF 3.2391 0.9146 0.1563 0.0146
40C/75%RH SA 3.1387 0.8764 0.1463 -0.0069

Table 4. Comparison of Standard Deviation of WVTR for Blister Types and Test Sites (Mean,
mg/cavity/day)
Standard Deviations (mg/cavity/day)
Condition Site PVC PVdC Ultrx Cold

23C/75%RH AB 0.0103 0.0307 0.0104 0.0160

23C/75%RH MK . 0.0010 0.0038 0.0007
23C/75%RH PF 0.0167 0.0078 0.0064 0.0154
23C/75%RH SA 0.0088 0.0006 0.0008 0.0111

40C/75%RH AB 0.0973 0.0137 0.0020 0.0240

40C/75%RH MK . 0.0161 0.0012 0.0058
40C/75%RH PF 0.0387 0.0102 0.0204 0.0218
40C/75%RH SA 0.0919 0.0091 0.0017 0.0032

9
 Figure 1

Figure 2

The above analysis uses a linear regression of the weights over time. A test that would just require using two
time points to determine the slope would be easier than running a regression. In the screening study, there were
8 time points for each blister type. Table 5 shows the mean slope across the containers using four different
slope calculations: 1) Slope using the 5th and 2nd time point, 2) slope using the 8th and 5th time point, 3) slope

10
using the 8th and second time point, and 4) slope using a linear regression. For the linear regressions, the initial
was left out of the analysis to remove any effect that may take place between the initial and the second time
point.

The effect of the curvature can be evaluated using this table by comparing the D5_2 which estimates the slope
from the 2nd to 5th time point to the D8_5 slopes which measures the slope from the 5th to 8th time point.
Notice that in general the slopes based on the 5th and 2nd time point are similar to the linear regression slope.
The differences between sites appear much smaller than differences between the blister types.

Table 5:
WVTR Blister Screening Study
Compare Blister Types and Test Sites
Slope’s Computed using Linear Regression
5th - 2nd, 8th - 5th, and 8th-2nd time points
(Initial removed)

Slope
Condition Site Method PVC PVdC Ultrx Cold

23C/75%RH AB D5_2 1.2391 0.2422 0.0319 0.0100

AB D8_5 1.1816 0.1260 0.0312 -0.0004
AB D8_2 1.2101 0.1841 0.0316 0.0048
AB reg 1.2153 0.1852 0.0290 0.0048
MK D5_2 . 0.1391 0.0290 0.0125
MK D8_5 . 0.1376 0.0272 -0.0043
MK D8_2 . 0.1383 0.0281 0.0041
MK reg . 0.1385 0.0282 0.0032
PF D5_2 1.2390 0.1419 0.0301 0.0082
PF D8_5 1.1946 0.1260 0.0258 -0.0122
PF D8_2 1.2168 0.1340 0.0280 -0.0020
PF reg 1.2234 0.1271 0.0279 -0.0011
SA D5_2 1.3270 0.1470 0.0295 0.0179
SA D8_5 1.2370 0.1365 0.0261 -0.0194
SA D8_2 1.2820 0.1417 0.0278 -0.0007
SA reg 1.2796 0.1415 0.0278 -0.0042

40C/75%RH AB D5_2 3.3583 0.8263 0.1379 0.0139

AB D8_5 2.6091 0.7567 0.1341 0.0056
AB D8_2 2.9732 0.7915 0.1360 0.0098
AB reg 2.9785 0.7889 0.1362 0.0109
MK D5_2 . 0.8472 0.1381 0.0089
MK D8_5 . 0.7375 0.1348 -0.0012
MK D8_2 . 0.7924 0.1364 0.0038
MK reg . 0.7913 0.1370 0.0032
PF D5_2 3.3216 0.9978 0.1638 0.0218
PF D8_5 3.0281 0.7580 0.1420 -0.0030
PF D8_2 3.1749 0.8779 0.1529 0.0094
PF reg 3.1827 0.8820 0.1517 0.0080
SA D5_2 3.3239 0.9245 0.1488 0.0096
SA D8_5 2.7048 0.7570 0.1394 -0.0187
SA D8_2 3.0144 0.8408 0.1441 -0.0046
SA reg 3.0186 0.8414 0.1442 -0.0073

While 23ºC/ 75% RH provides the best storage condition for PVC blisters, 40ºC/ 75% RH was selected as the
preferred storage condition based on reduced variability of WVTR of the barrier blisters as measured by RSD.
PVC blisters saturate relatively quickly at 40ºC/ 75% RH, and use of the initial weight and day 2 weight to

11
calculate WVTR provided the highest slope and the best estimate of WVTR compared with, for example,
WVTR calculated using the day 2 to day 4 weights as shown in Table 6.

Table 6:
WVTR Blister Screening Study
Compare PVC Slopes (Day 2 - Initial) vs (Day 4 - Day 2)

Slope
Condition Site Day 2 - Initial Day 4 - Day 2

23C/75%RH AB 1.317 1.218

PF 1.327 1.185
SA 1.371 1.368

40C/75%RH AB 3.982 3.526

PF 3.747 3.348
SA 3.932 3.544

Use of the 40ºC/ 75% RH storage condition also allows manufacturer’s to correlate results with stability data
generated at the 40ºC/ 75% RH ICH storage condition.

Recommendations
Based on the results of this study, we recommend:
 The use of 40ºC/ 75% RH storage condition based on reduced variability, as measured by RSD, and
ability to correlate with an ICH stability test condition.
 The use of the ASTM E 96 calculator to determine a constant partial moisture vapor pressure
differential, the number of blisters and balance sensitivity to ensure data precision and accuracy.
 A test duration of 2 days for PVC and 35 days for barrier blisters at 40ºC/ 75% RH.
 The use of 2 time points for PVC at initial and day 2 and the use of 5 (weekly) time points for barrier
blisters beginning at day 7. Eliminate the initial time point (time ‘0’) for barrier blisters.
 Eliminate the use of empty controls when using non-paper backed lidding foil.
 The use of linear regression for calculating WVTR for barrier blisters.
 Report results in mg/day/blister cavity and eliminate reference to blister classification to enable
comparison with other container closure systems (e.g. bottles).

The recommendations are made with pharmaceutical manufacturer’s needs and capabilities in mind to provide a
platform for meaningful comparison between container closure systems, and to enable determination of a
WVTR performance space in which the stability of a solid oral dose form can be ensured.

12
 Appendix 1
Reason for use of E- 96 Calculator

During design of the blister study, it was necessary to account for a number of
simultaneous random variables. The following questions arose:

1. How to assure that there was enough desiccant in the package to carry the test to
completion?
2. How to assure that steady state would be reached within the time of the test?
3. How to verify that steady state had been reached?
4. How to minimize variability resulting from limited sensitivity of the balance
used?
5. What should be the duration of the test?
6. What weighing intervals should be used?
7. What desiccant should be used?

Each attempt to fix any one of the variables above resulted in uncertainty about the
others. It was obvious that the questions were inter-related, but the solution was not
obvious.

A literature review on the subject of steady state water vapor transmission revealed an
ASTM method that provided an answer. The method is ASTM E96 – 00, “Standard Test
Methods for Water Vapor Transmission of Materials”. The method provides for control
of the following test parameters, calculations and analyses:

1. Balance sensitivity < 1% of weight change during the period of steady state
2. Weights shall be accurate to 1% of weight change during the steady state period.
3. Weigh often enough to provide 8 or 10 data points during the test
4. Record time to 1% of time span between successive weighings.
5. Terminate the test if the total change in weight exceeds 10% of the initial weight
of the desiccant.
6. The method specifies anhydrous calcium chloride, dried at 400 F (200 C), as a
desiccant
7. The method specifies use of a blank for low transmission rates
8. Periodic weight change equals or exceeds 20% of (100 x scale sensitivity)
9. For graphic analysis, plot weight against time. Six points determines the straight
line fit to determine the slope. Or,
10. For numerical analysis, use least squares regression analysis of weight as a
function of time.

To design this experiment required an estimation of the value of water vapor permeation
for the materials to be tested. This was done using values obtained during previous
measurements of water vapor permeability.

Four steps were taken.

13
 1. An iterative program was written in Virtual Basic to calculate outcomes for
various scenarios among the parameters. The program used the guidelines
provided in E-96 for the variable parameters.
2. Silica gel was chosen as the preferred desiccant for blisters. This choice was
made on the basis of previous laboratory experience with several desiccants,
including silica gel.
3. The estimated WVTRs were used to determine the expected uptake of water.
This amount was compared with the water sorption isotherm for silica gel to
establish the minimum amount of silica gel required to avoid saturation of the
desiccant before the end of the experiment. It was determined that the amount of
desiccant in the package must be enough to maintain RH of not more than 10%
within the headspace of the package.
4. Test duration was chosen to allow for 8 data points, including zero time.

To ensure linearity of the weight gain vs. time data, one needs to keep the driving force
for water-vapor permeation as nearly constant as possible during the testing period.
Since the external water activity (i.e. RH) is maintained constant at 75% at either test
temperature, this means that one would aim to keep the internal water activity as nearly
constant as is practically feasible. We recommend placing sufficient desiccant in the
container to ensure that the internal headspace RH does not exceed 10%. This way, the
driving force would stay reasonably constant, between 75% and 65% RH, throughout the
testing period.

Literature search suggests that a typical silica gel desiccant absorbs about 7% to 10% by
mass of water at 40ºC/10% RH. Thus, in the proposed test method, one must place
sufficient amount of desiccant in each cavity such that the total cumulative mass increase
per mass of desiccant stays below 7% to 10% wt. To be on the conservative side, we
chose 7% for our recommendation. The following table was generated to help determine
the minimum amount of desiccant needed in each blister cavity according to the expected
WVTR range of the material under test.

Anticipated
Minimum
Maximum Maximum Amount Target Maximum Amount of Moisture in
Recommended Amount of
WVTR, of Moisture Desiccant at End of Test, wt%
Test Duration, Desiccant
mg/cavity/day Entering the (corresponding to ~10% RH in the
Days Per Cavity,
at 40C/75% Cavity, mg Moisture Sorption Isotherm of Silica Gel)
mg
RH
0.1 35 3.5 7% 50
0.5 35 17.5 7% 250
1 35 35 7% 500
4 2 8 7% 114

Appendix 2 shows the results of calculations for amount of desiccant, weighing

intervals and duration of test applied for the experiment reported here.

14
 Appendix 2
PQRI Experiment for WVTR Test Method
Unit Dose Packages (Blisters)
Test Intervals, and Weight Gain
ASTM E 96 Calculator Information

(Size 0) PVC blisters,

Desiccant: Silica gel
Balance Sensitivity 0.1 mg, Conditions 23 C/75% RH
Estimated WVTR= 0.5 mg/day/cavity at 23 C/75% RH
Sealed blisters
Duration of Test, days 14

Weighing Interval, days 2 (total 7

points)
Estimated Wt Gain/Interval/10- 10 mg
cavity, mg
Amount Desiccant 1
Required/cavity, g
No of Blister Cards 1(10)
(cavity/card) for each weighing
Number of Replicate Weights 10
Total Number of Blister Cards 40
for 4 labs

Balance Sensitivity 0.1 mg, Conditions 40C/75% RH

Estimated WVTR= 1.2 mg/day/unit at 40 C/75% RH,
Sealed blisters
Duration of Test, days 14
Weighing Interval, days 2 (total 7
points)
Estimated Wt Gain/Interval/10- 24 mg
cavity, mg
Amount Desiccant 1
Required/cavity, g
No of Blister Cards 1(10)
(cavity/card) for each weighing
Number of Replicate Weights 10
Total Number of Blister Cards 40
for 4 labs

15
 (Size 0) PVdC /60 gms blisters,
Desiccant: Silica gel
Balance Sensitivity 0.1 mg Conditions 23 C/75% RH
Estimated WVTR= 0.065mg/day/cavity at 23 C/75% RH
Sealed blisters
Duration of Test, days 49

Weighing Interval, days 7 (total 7

points)
Estimated Wt Gain/Interval/20- 9.1
cavity, mg
Amount Desiccant 1
Required/cavity, g
No of Blister Cards 2(10)
(cavity/card) for each weighing
Number of Replicate Weights 10
Total Number of Blister Cards 80
for 4 labs

Balance Sensitivity 0.1 mg Conditions 40C/75% RH

Estimated WVTR= 0.47mg/day/unit at 40 C/75% RH,
Sealed blisters
Duration of Test, days 49
Weighing Interval, days 7 (total 7
points)
Estimated Wt Gain/Interval/10- 32.9
cavity, mg
Amount Desiccant 1
Required/cavity, g
No of Blister Cards 1(10)
(cavity/card) for each weighing
Number of Replicate Weights 10
Total Number of Blister Cards 40
for 4 labs

16
 (Size 0) Ultrx 2000 blisters,
Desiccant: Silica gel
Balance Sensitivity 0.1 mg Conditions 23 C/75% RH
Estimated WVTR= 0.018mg/day/unit at 23 C/75% RH
Sealed blisters
Duration of Test, days 98

Weighing Interval, days 14 (total 7

points)
Estimated Wt Gain/Interval/30- 7.6
cavity, mg
Amount Desiccant 1
Required/cavity, g
No of Blister Cards 3(10)
(cavity/card) for each weighing
Number of Replicate Weights 10
Total Number of Blister Cards 120
for 4 labs

Balance Sensitivity 0.1 mg Conditions 40C/75% RH

Estimated WVTR= 0.085mg/day/unit at 40 C/75% RH,

Duration of Test, days 98

Weighing Interval, days 14 (total 7
points)
Estimated Wt Gain/Interval/10- 11.9
cavity, mg
Amount Desiccant 1
Required/cavity, g
No of Blister Cards 1(10)
(cavity/card) for weighing
Number of Replicate Weights 10
Total Number of Blister Cards 40
for 4 labs

17
 (Size 0) Coldform Alu/alu blisters,
Desiccant: Silica gel
Balance Sensitivity 0.1 mg Conditions 23 C/75% RH
Estimated WVTR< 0.01mg/day/unit at 23 C/75% RH
Sealed blisters
Duration of Test, days 98

Weighing Interval, days 14 (total 7

points)
Estimated Wt Gain/Interval/30- <1
cavity, mg
Amount Desiccant 0.6
Required/cavity, g
No of Blister Cards 5(6)
(cavity/card) for each weighing
Number of Replicate Weights 10
Total Number of Blister Cards 200
for 4 labs

Balance Sensitivity 0.1 mg Conditions 40C/75% RH

Estimated WVTR<0.01mg/day/unit at 40 C/75% RH,

Duration of Test, days 98

Weighing Interval, days 14 (total 7
points)
Estimated Wt Gain/Interval/30- <1
cavity, mg
Amount Desiccant 0.6
Required/cavity, g
No of Blister Cards 5(6)
(cavity/card) for each weighing
Number of Replicate Weights 10
Total Number of Blister Cards 200
for 4 labs

1
Barry J, Bergum J, Chen Y, Chern R, Hollander R Klein D, Lockhart H, Malinowski D, McManus R,
Moreton C, Mueller A, Nottingham L, Okeke C, O’Reilly D, Rinesmith K, and Shorts S, (PQRI Container-
Closure Working Group), Basis for using Moisture Vapor Transmission Rate pre Unit Product in the
evaluation of Moisture-Barrier Equivalence of Primary Packages for Solid Oral Dosage Forms,
Pharmacopeial Forum, (2005), 31, (1) Jan.—Feb., 226 – 269.

18