Quantifying the Vial-Capping Process: Reexamination Using ... · 2.6.VialLandSealStudies Serum...

10.5731/pdajpst.2019.010363Access the most recent version at doi: 171-18474, 2020 PDA J Pharm Sci and Tech

Robert Ovadia, Philippe Lam, Vassia Tegoulia, et al. Micro-Computed TomographyQuantifying the Vial-Capping Process: Reexamination Using

on June 22, 2020Downloaded from on June 22, 2020Downloaded from

RESEARCH

Quantifying the Vial-Capping Process: Reexamination UsingMicro-Computed Tomography

ROBERT OVADIA, PHILIPPE LAM, VASSIATEGOULIA, and YUH-FUNMAA*

Pharmaceutical Processing and Technology Development, Genentech, A Member of the Roche Group, San Francisco, CAUSA © PDA, Inc. 2020

ABSTRACT: A vial-capping process for lyophilization stopper configurations was previously quantified using residual

seal force (RSF). A correlation between RSF and container closure integrity (CCI) was established, and component

positional offsets were identified to be the primary source of variability in RSF measurements.

To gain insight into the effects of stopper geometry on CCI, serum stoppers with the same rubber formulation were

investigated in this study. Unlike lyophilization stoppers that passed CCI (per helium leak testing) even with RSF of

0N owing to their excellent valve seal, serum stoppers consistently failed CCI when RSF was <15.8 N. When the plug

was removed, both types of stoppers exhibited a comparable critical lower RSF limit (19–20N), below which CCI could

not be maintained. When CCI was retested at later time points (up to 6 mo), some previously failed vials passed CCI,

suggesting that CCI improvement might be related to rubber relaxation (viscous flow), which can fill minor imperfec-

tions on the vial finish.

To confirm component positional offsets are the primary sources of RSF variability, a novel quantification tool—micro-com-

puted tomography (micro-CT)—was used in this study. Micro-CT provided images for quantification of positional offsets of

the cap and stopper that directly correlated with RSF fluctuations. Serum stoppers and lyophilization stoppers are comparable

in RSF variations, although lyophilization stoppers are more robust in CCI. The use of micro-CT provides a nondestructive and

innovative tool in quantitatively analyzing component features of capped vials that would otherwise be difficult to investigate.

KEYWORDS: Micro-computed tomography, Micro-CT, Residual seal force, RSF, Vial capping, Container closure in-

tegrity, CCI, Primary packaging components, Helium leakage, Vial, Stopper, Crimp cap.

1. Introduction

A robust container closure system (CCS) is the final

defense to sterile products. Integral CCSs protect par-

enteral drug products from environmental influences

such as microbial contamination or the penetration of

extraneous substances that may be detrimental to prod-

uct quality (1). Furthermore, lyophilized products,

which are often packaged under reduced pressure, and

sensitive products, which are packaged under inert

atmosphere, require gas-tight closures. Vial CCSs typi-

cally contain three primary packaging components: a

glass or plastic vial, a rubber stopper, and an aluminum

crimp cap. The fit between these components influences

the integrity of the seal, which can be qualitatively and

quantitatively assessed. Qualitative methods typically

apply visualization tools to characterize critical sealing

surfaces between a stopper and a vial (2). These tools

are valuable for component screening and troubleshoot-

ing existing CCSs. There are many quantitative test

methods that evaluate container closure integrity (CCI)

and they are thoroughly described in USP Chapter

<1207> and in various publications (3, 4), but these

methods are generally not effective as process develop-

ment tools. Any of these quantitative methods can illus-

trate only some aspects of the CCI and need to be

combined with other orthogonal qualitative and quanti-

tative methods to gain the full understanding of the

CCS. For example, a method listed in USP <1207>

may clearly determine the pass/fail outcome, but it may

not be able to demonstrate the seal robustness (valve or

land) of the CCI.

Residual seal force (RSF) represents a quantitative

method for characterizing seal quality, and it is

* Corresponding Author: Yuh-Fun Maa, Genentech, A

Member of the Roche Group, 1 DNA Way, South San

Francisco, CA 94080; e-mail: [email protected]

doi: 10.5731/pdajpst.2019.010363

Vol. 74, No. 2, March--April 2020 171

on June 22, 2020Downloaded from

considered a valuable tool for critical capping parame-

ter identification to ensure the consistency and stand-

ardization of the capping process (5–8). The RSF

methodology has been extensively studied in recent

years (9–13). We previously evaluated manufacturing-

scale RSF data in which the capping process was quan-

tified for lyophilization stopper configurations by

establishing a correlation between RSF and CCI (14).

In that study, we also observed highly variable RSF

measurements, which hindered meaningful RSF limits

to be set for capping operations. Although several fac-

tors might contribute to RSF variation, the intrinsic

variability in dimensional tolerances of each primary

packaging component was considered to be the primary

root cause based on indirect evidence. This current

study applied a similar approach to assessing serum

stopper configurations and was extended to evaluate

the durability of CCI during shelf life. Because both

the lyophilization and serum stoppers are of the same

rubber formulation, any observed differences in the

relationship between RSF and CCI can mostly be

attributed to geometric differences between the two

styles of stoppers.

The primary objective of this study was to provide

direct visual evidence on the relationship between the

primary packaging component’s intrinsic dimensional

tolerances and RSF variability using micro-computed

tomography (micro-CT). Micro-CT is a nondestructive

method capable of analyzing an intact CCS including

both glass and plastic vials. Micro-CT was previously

used by Mathaes et al. to visualize 1) crimping defects,

2) aluminum cap and flip-off button geometries of dif-

ferent CCSs, and 3) unsymmetrically capped vials

(10–12). Observing unsymmetrically capped vials led

to the possibility of quantifying the capping process,

and recent technological advances produced somewhat

affordable benchtop micro-CT instruments with high

image resolution capability that could be used as a

quantitative tool. Haeuser et al. applied micro-CT to

characterize the cake structure of lyophilized drug

products (15). Hindelang et al. demonstrated the utility

of micro-CT to measure vial wall thicknesses (16).

Overall, micro-CT is not a high-throughput method,

and it requires complex image analysis; however, it is

the only tool available to visualize and quantify inter-

nal features of an intact CCS at micrometer resolution.

In this study we used micro-CT to extract component

positional offsets (i.e., how well the components are

centered relative to each other), which might be related

to inherent component variations.

2. Materials and Methods

2.1. Primary Packaging Components

The primary packaging components used in this study

included 20-mL type 1 glass vials (Schott Schweiz AG),

20-mm D777-1 serum rubber stoppers (Daikyo Seiko) in

a ready-to-sterilize format, and 20-mm aluminum seals

with plastic flip-off buttons (West Pharmaceutical Serv-

ices, Inc.). The vials were washed and depyrogenated

before use; all other components were used as received.

These are custom materials that have slightly different

dimensions and tighter tolerances (not disclosed) than

commercially available, off-the-shelf components based

on International Organization for Standardization specifi-

cations. These components have been developed, refined,

and eventually used in multiple commercial products over

many years. All components are extensively tested before

use as part of our raw material release process to ensure

specification compliance.

2.2. Vial Capping

Vials were capped using an Integra Laboratory Crimper

(Genesis Packaging Technologies). The primary capping

parameters tested in this study included capping precom-

pression force and capping plate-plunger distance. Other

capping parameters, such as capping plate geometry and

angle, capping plate travel distance, and rotational speed of

the plates, were held constant. The definition and function

of these capping parameters were previously described (10).

2.3. RSF Measurements

RSF was measured using an automated Residual Seal

Force Tester (Genesis Packaging Technologies). Test-

ing was conducted as per a previously described proto-

col (14). All RSF measurements were performed

between 24 h and 30 d postcapping using the lowest

(111N) force level; previous data (14) suggest that

there is no substantial change in RSF values with

respect to time for the tested configuration.

2.4. Vial and RSF Tester Orientation Study

The experiment was conducted identically to the previous

study of lyophilization stoppers for direct comparison (14).

In short, 40 vials were labeled, capped, and randomly di-

vided into two even groups, Group 1 and Group 2. The

vials were capped with an RSF target of 50N. All capped

172 PDA Journal of Pharmaceutical Science and Technology


vials were held for 24 h before the first RSF measurement.

During each measurement, vials in Group 1 were ran-

domly oriented on the RSF tester base plate, whereas the

orientation of vials in Group 2 was fixed.

2.5. Helium Leak Testing

Our previously published procedure (14) was applied in

this study. Briefly, using a diamond cutting blade, a slit

was cut at the heel of the vial to allow helium gas purging

into the vial. Potential leaks from the capped area could

be detected using an ASM340 mass spectrometric helium

leak (He-leak) detector (Pfeifer Vacuum) equipped with a

custom vial holder. A leak rate of ≥1.0� 10�7 mbar L/swas defined as a CCI failure. Unless otherwise noted, He-leak measurements were performed within 1 h of capping.

2.6. Vial Land Seal Studies

Serum stoppers were mounted in a custom fixture, and

they were cut using a sharp blade to separate the plug

from the flange. The flange was positioned at the top

center of the vial for capping. Refer to our previous

study for details (14).

2.7. Micro-CT Scanning

Micro-CT provides the location of internal features of

an object that would otherwise be difficult to visualize

in a nondestructive method. The entire scanning/imaging

process is schematically presented in Figure 1. X-rays

emitted by the source pass through a sample, and the

transmitted rays are detected and recorded to generate a

projection of the sample based on its X-ray absorption. A

series of projections are captured consecutively by rotat-

ing the sample by a fraction of a degree, until a full 180˚

or 360˚ rotation is achieved. After scanning, images are

reconstructed by transforming the data into a stack of

cross-sections, which are imported into image processing

software for the construction of a three-dimensional (3D)

image. A more detailed description of the scanning,

reconstruction, and imaging sequence is provided below.

2.7.1. Scan Image Acquisition and Image Re-

construction: Vials for scanning were obtained from

the orientation study by selecting the five least and

most variable vials from Group 1 (randomly oriented).

A SkyScan 1272 X-ray microtomograph (Bruker

MicroCT) was used in this study. Parameter setting and

description for both scan image acquisition and image

reconstruction are summarized in Table I. Because the

glass and the aluminum have high X-ray attenuation

coefficients (i.e., the fraction of X-ray beams that are

absorbed or scattered per unit of thickness of material),

source voltage, source current, and the filter were

adjusted to increase the contrast between the three dif-

ferent materials of the primary packaging components:

glass, rubber, and aluminum. The total scan time per vial

Figure 1

Micro-CT scanning, reconstruction, and image processing schematic.



TABLE I

Selected Scan Image Acquisition and Reconstruction Parameters

Parameter Set Value Parameter Description

Scan image

acquisition

Energy source

(voltage,

current)

100 kV, 100 lA Source voltage and current are adjusted for each sample

depending on its material. High absorbing materials,

such as glass and aluminum, need high energy.

X-ray filter Al 0.5mm +

Cu 0.038mm

Filter placed between the X-ray source and the sample

reduces the polychromaticity of the source (filtering out

X-ray energies outside the target); strong filters are

needed for high absorbing materials.

Pixel setting and

resolution

2k (1632� 1092)

pixel setting with

a length of 15 lm

Pixel length can be adjusted by increasing the pixel

setting (1k, 2k, and 4k), moving the sample, or both.

Smaller pixel lengths generate images of higher

resolution.

Exposure time 2575ms Exposure time is affected by camera position, source

voltage and current, and selected filter. Higher exposure

times are generally needed for higher absorbing

materials; however, the detector may deteriorate faster

under long high-power scans.

Rotation step 0.3˚ up to 180˚ A projection image is generated at each rotation step.

Generally, smaller rotation steps produce higher quality

projections.

Frame

averaging

7 frames Multiple projections are averaged at each rotation step to

reduce background noise.

Vertical random

movement

100 lm The sample is moved up and down between projections

at random distances within 100 lm to reduce ring

artefacts (described below).

Image

reconstruction

Beam hardening 45% Beam hardening artefacts cause the edge of an object of

the same material to be brighter than the center. A

procedure of postcorrection during reconstruction

minimizes these artefacts. Postcorrection values are

held constant between scans if the material and scan

settings are identical.

Gaussian

smoothing

1 Postcorrection smoothing can reduce background noises

and is held constant between scans if the material and

scan settings are identical.

Ring artifact

correction

2–4 Ring artefacts, commonly caused by dust or

miscalibrated detector elements (19), appear as

concentric circles in a reconstructed slice. Their effect

is reduced by applying a reduction value between 0 (not

corrected) and 20 (heavily corrected).

Misalignment

compensation

Variable Misalignment compensation values are visually assessed

for the best alignment within a reconstructed slice.

Reconstructed

cross-sectional

images

550 The number of cross-sectional images is kept constant at

550; these images encompass the top of the aluminum

cap (plastic button removed) and the bottom of the vial

flange.



was approximately 4 h. Note that high source voltage

combined with long scan times may deteriorate the detec-

tor more rapidly and may cause detector artefacts (tempo-

rary “bleaching”). The scanning parameters selected for

this study were optimized for our application. Projections

were reconstructed using NRecon software (Bruker),

where parameters (Table I) were set to minimize artefacts

(beam hardening and ring artefacts) and background

noises (Gaussian smoothing), as well as optimize image

construction (misalignment compensation).

2.7.2. Analysis: For each vial, a stack of reconstructed

images was analyzed using the Simpleware ScanIP soft-

ware (Synposys Inc.). A consistent set of image-process-

ing operations, including segmentation, morphological,

and smoothing, was applied to visualize each component

independently. ScanIP was used to fit circles (using

bounds method) to both the stopper and the cap at the

same cross-sectional plane, at approximately the midpoint

of the stopper flange thickness for each vial.

3. Results and Discussion

3.1. Quantification of Serum Stopper Configurations

We previously demonstrated a methodology to quan-

tify the vial-capping process for lyophilization stopper

configurations via correlating RSF to CCI (14). The

RSF-CCI relationship would provide scientific justifi-

cations for setting acceptable RSF limits, which would

mitigate the risk of failing CCI. This current study

focused on serum stoppers. Figure 2 allows a visual

comparison of the lyophilization stopper (Figure 2a)

and the serum stopper (Figure 2b), and their key simi-

larities and differences are summarized in Table II.

The difference in plug design (feature 6 in Table II)

may have a major impact on CCI. Capping and CCI

performance of the two stopper types were compared

in this study.

3.1.1. Relationship between RSF and CCI: To assess

the capability of serum stoppers and their impact on

CCI, 60 vials were capped at RSFs ranging from 4.0 N

to 30.6 N and tested for CCI by the He-leak method.

Twenty-seven vials failed He-leak testing (i.e., having

a leak rate of >1.0� 10�7 mbar L/s) as shown in a box

plot (Figure 3a). Vials with greater RSFs are more

likely to pass the He-leak test (Figure 3b); all vials

with an RSF of >15.8 N passed He-leak testing,

whereas all vials of <9.6 N failed. The pass/fail pattern

is unpredictable for vials with intermediate RSF values.

These data establish a quantitative correlation between

RSF and CCI and allow 15.8 N to be proposed for fur-

ther statistical analysis to calculate the lower RSF

limit, which is an important parameter for good manu-

facturing practice (GMP) manufacturing capper setup.

The proposed lower RSF limit is specific to this CCS

and is not intended to cover other CCSs.

These results were not consistent with a similar study

performed by Mathaes et al. (11). In that study, no CCI

failures were observed for serum stopper configurations

capped at low RSFs. An obvious difference between the

two studies is the use of different component (vial, stop-

per, and cap) batches. To validate this lower RSF limit

for commercial manufacturing use, a robustness study

involving a much larger dataset (including multiple

component lots) should be performed.

3.1.2. Effect of Land Seals on RSF and CCI: The land

seal of the serum stopper was isolated by cutting off

the plug. Sixty vials were capped with plug-free serum

stoppers, and their RSFs ranged from 7.2 N to 25.3 N.

He-leak testing results showed that 28 vials failed the

test, whereas 32 passed (see the box plot in Figure 4a).

Vials with lower RSFs have a higher risk of failing the

He-leak test; all vials with RSF of <15.4 N failed CCI,

whereas all vials with RSF of >19.8 N passed (Figure

4b). This trend is comparable with that of the intact se-

rum stoppers, only with a slightly higher lower RSF

limit, 19.8 N versus 15.8 N (Table III). These compara-

ble lower RSF limits for the intact and plug-free serum

stopper suggest that the plug of the serum stopper

Figure 2

Comparison between (a) lyophilization stopper and

(b) serum stopper, with numbered features corre-

sponding to the similarities and differences listed in

Table II.



TABLEII

Sim

ilaritiesandDifferencesbetw

eentheSerum

andLyophilizationStoppers

Feature

Number

aSerum

Lyophilization

Impact

onCCI

Rationale

Key

similarities

1Stopper

form

ulation

Major

Stopper

form

ulationandouterdiameter

oftheplugcan

potentially

impactCCI,so

they

werekeptconstantfor

thiscomparison.

2Outerdiameter

oftheplug

Major

Key

differences

3Flangethicknessover

thevial

finishisslightlythicker

than

thatofthelyophilization

stopper

byabout3%

Flangethicknessover

thevial

finishisslightlythinner

than

thatoftheserum

stopper

by

about3%

Minor

Theflangethicknessdifference

betweentheserum

andlyophilizationstoppersisnotexpectedto

be

significantbecause

thisdifference

isminorand

possibly

within

dim

ensiontolerances.

4Dim

pledtopcenterofflange

Flatflange

Minor

Flangethicknessover

thevialfinish,as

notedin

#3,

iscomparable.

5Nofluoropolymer

film

ontopof

flange

Fluoropolymer

film

presenton

topofflange

Minor/

none

Fluoropolymer

film

isverythinandprovides

smoothnesstothetopofthestopper;nofilm

ison

thestopper

landseal.

6Pluggeometry:

Concavefeature

insideplug,

minim

alstructuralsupport

Fluoropolymer

film

placement

resultsinshorter

valvesealof

0.71mm

Pluggeometry:

Ventandlegsprovide

structure

Fluoropolymer

film

placementresultsin

longer

valvesealof0.95mm

Major

Lyophilizationstopper

plugmay

form

bettervalve

sealwithglasscompared

withserum

stopper

plug

owingto

structuralsupportfromthelyophilization

stopper

andplacementoffluoropolymer

film

.

aRefer

toFigure

2foracomparisonofthenumbered

featuresoftheserum

andlyophilizationstoppers.



provides minimal protection in the event of a compro-

mised land seal.

3.1.3. Quantitative Comparison between Serum and

Lyophilization Stoppers: A similar set of studies was

previously performed using lyophilization stoppers

(14), and the data for both stoppers types are compared

in Table III. Previously, all 20 intact lyophilization

stoppers passed CCI under RSFs as low as 5N [see Fig-

ure 9 in Ovadia et al. (14)]. That study also tested 20 ly-

ophilization stoppers whose flanges had been removed

(i.e., flange-free) (14). The stoppers had no land seal,

so they could not register any measurable RSF. All 20

vials passed He-leak testing; thus, it implied that the

valve seal of the lyophilization stopper is essential in

maintaining CCI before capping, which is contrary to

the serum stopper. Note that both stoppers have the

same outer diameter, suggesting that it is the overall

design of the plug, not just the plug diameter, that plays

an important role in maintaining CCI (feature 6 in Ta-

ble II). Overall, the lyophilization stopper outperforms

the serum stopper in maintaining CCI when both have

comparable RSF, particularly in the low range of the

RSF.

Compared with the plug-free serum stoppers of this

study, the lyophilization stoppers of the previous study

(14) displayed a similar RSF to CCI trend. The plug-

free lyophilization stoppers could pass CCI only if their

RSF values were >20.1 N (14). With a similar lower

RSF limit, that is, 19.8 N versus 20.1 N, these two stop-

pers (serum and lyophilization) proved to be compara-

ble if their plug was absent. This is not surprising

because both stoppers have the same rubber formula-

tion and similar flange thickness over the vial finish

(features 1 and 3 in Table II and Figure 2a and b). As

pointed out by Morton and Lordi (7, 8), the stopper

must be compressed with sufficient force by the crimp

Figure 3

(a) Box plot and (b) scatter plot of RSF values for serum stoppers. All vials with an RSF >15.8 N passed He-

leak testing.

Figure 4

(a) Box plot and (b) scatter plot of RSF values for plug-free serum stoppers. All vials with an RSF >19.8 Npassed He-leak testing.



cap to properly engage the large flange sealing surface;

otherwise, CCI will not be robust.

3.1.4. Consideration on Using Lyophilization Stop-

pers for Liquid Products: Lyophilization stoppers are

primarily used for freeze-dried products. With the find-

ing above justifying a much more robust CCI capabil-

ity, it may be beneficial to adapt the design of the

serum plug or use lyophilization stoppers for liquid

products. These findings are specific to the testing con-

figuration. Certainly, development scientists and engi-

neers need to consider the overall pros and cons of

using either stopper with any specific product. Potential

advantages of using lyophilization stoppers on non-

lyophilized (i.e., liquid) products include 1) maintain-

ing CCI over a wider range of RSFs, 2) reducing the

number of change parts for manufacturing, and 3) sim-

plifying material management. However, this approach

does come with drawbacks: 1) higher costs associated

with the lyophilization stopper, 2) needing to refill the

stopper bowl more frequently during manufacturing

owing to the larger volume of the lyophilization stop-

per, and 3) greater product interactions because of

higher stopper surface area exposed to the drug product

solution. In the end, a balanced decision weighing tech-

nical and business considerations is needed when

adopting this approach.

3.2. Impact of Shelf Life on CCI

Rubber stoppers normally exhibit viscoelastic charac-

teristics and may undergo viscous flow, particularly

under stress. This property may affect CCI during stor-

age. The durability of CCI during the shelf life (storage

at 226 2˚C) of vials capped with plug-free serum stop-

pers was evaluated. To compare with the CCI data in

Figure 4b, in which He-leak testing was performed

within 1 h postcapping, the same vials were tested at 1,

3, 7, 28, and 180 d postcapping (Figure 5a). CCI

improvement was observed during the first 7 d; the

number of originally failed vials decreased from 28 to

19. No further improvement was observed after day 7

up to 180 d. The nine vials (Figure 5b, blue triangles)

that went from failing CCI to passing were those origi-

nally with the highest RSF values on the curve in Fig-

ure 5b. As a result, the lower RSF limit decreased from

19.8 N to 13.7 N.

TABLE III

Quantitative Comparison, RSF to CCI Relationship, between Serum and Lyophilization Stoppers

Serum RSF Failure Limit Lyophilization RSF Failure Limita

Intact stopper 15.8N 0N (No observed CCI failures)

Plug-free stopper 19.8N 20.1N

Flange-free stopper Not tested Not applicable; no flange (No observed CCI failures)aAs determined previously (14).

Figure 5

(a) Scatter plot displaying impact of time when performing CCI testing. Some vials that failed the 1-h time

point passed at later time points. No change is seen between 7 and 180 d. (b) Scatter plot of RSF values for

plug-free serum stoppers with vials that changed from failing to passing (blue triangles) when tested at 7 d.



The trend in Figure 5a and b is not expected to change

beyond 180 d. The improvement may be related to rub-

ber relaxation (or viscous flow), which allows the rub-

ber to fill minor imperfections on a vial finish,

particularly during the first 7 d. Viscous flow occurs

more readily under stressed conditions, and the flow

rate increases with higher force. This may explain why

the remaining 19 vials still failed CCI; the stoppers on

these vials had the lowest RSFs and were not causing

sufficient viscous flow. This trend should be configura-

tion specific. In addition, although rubber relaxation

can improve CCI stability, it should not be used to jus-

tify the acceptance of vials with marginal RSF values

immediately after vial capping in anticipation that

these vials would “self-seal” after some time. The goal

of the manufacturing-scale capping process is always

to achieve immediate 100% CCI. This study stresses

the importance of measuring CCI under worst-case

conditions, including performing RSF promptly after

capping. However, the data presented above do demon-

strate that a container with an integral seal will retain

that state, provided the storage temperature is not too

extreme (17, 18). This observation is of significance

because regulatory authorities typically expect proof

that CCI will be maintained throughout the product’s

shelf life.

3.3. RSF Variability Caused by Inherent Component

Variations by Indirect Method

Highly variable RSF measurements were previously

observed for lyophilization stopper configurations and

were attributed to inherent tolerances of the primary

packaging components (14). This study was repeated

for serum stopper configurations. To focus on the effect

of vial orientation on the RSF tester, vials were sepa-

rated into two groups to be capped in random (Group

1) and fixed (Group 2) orientations. Each group con-

tained 20 vials, and the RSF of each vial was measured

20 times. Figure 6a and b summarize these results as

box plots for the randomly oriented vials and those of

fixed orientation, respectively. The %RSD of each vial

in their respective group is summarized as a box plot in

Figure 6c. The average %RSD (highlighted by the red

line in Figure 6c) for the randomly oriented group

(9.2%) was two times greater than that of the group of

fixed orientation (4.6%).

This observation is consistent with the previous study

(14) of lyophilization stopper configurations, which

found that RSF values of randomly oriented vials

display higher intravial variability relative to those of

the vials of fixed orientation. We previously implied in-

herent component tolerances as the main contributor to

RSF variability but lacked direct evidence. Examples of

inherent tolerances may include component dimensions,

flatness and uniformity of the sealing surfaces, shape

uniformity of the aluminum skirt, as well as the posi-

tional offsets of these components when capped. This

current study offered direct evidence by applying micro-

CT for quantitative analysis of positional offsets.

3.4. Micro-CT Scanning

Inherent tolerances of vial components may lead to off-

center assemblies (or positional offsets) of the vial, the

stopper, and the cap. We speculated that capped vials

with less RSF variability were more centered, whereas

capped vials with larger RSF variability were less

centered (i.e., greater positional offset). To verify this

hypothesis, vials were selected from the randomly ori-

ented group (Group 1, see Section 3.3), including five

vials with most variable RSF (Vials 5, 7, 13, 15, and

20) and five vials with least variable RSF (Vials 1, 3, 8,

12, and 14) (refer to Figure 6b). These vials were tested

for positional offsets using micro-CT.

3.4.1. Development of Micro-CT Scan Acquisition

and Image Reconstruction Settings: Best micro-CT

practices suggest that scan acquisition parameters

should be tuned for the shortest scan time that would

still provide the user with good image quality and reso-

lution for the features of interest. In this study, the fea-

tures of interest are the exterior boundaries of the cap

and the stopper. In the early phase of scan parameter

setting, a vial was scanned for 24 h using enhanced pa-

rameters, such as longer scanning times compared

with what is listed in Table I. Although enhanced pa-

rameters produced better image quality, they were

not ideal for the detector. For example, long scan

times would cause permanent damage (long-term

effect) to the detector and would also result in

bleaching artefacts to images taken subsequently

(short-term effect). An iteration of scan parameter

tweaking was performed to minimize scan time while

still providing the same quantitative results. Figure

7a and b display images of the same vial scanned

using enhanced and selected scanning parameters,

respectively, for comparison. The image generated

from enhanced scanning parameters is of higher re-

solution with fewer artefacts compared with the

counterpart from selected parameters. Nevertheless,



key features remained highly visible from the image

derived from selected parameters and had no impact

on data interpretation. Thus, these selected parame-

ters were used for subsequent scanning.

Reconstruction parameters were optimized to refine

scanned images. Ring artefacts, for example, were

common in scanned images (see Figure 8) in this

study. Although visually distracting, ring artefacts did

not directly affect analysis because they were located

away from the features of interest for our case. The

effort of decreasing ring artefacts by applying ring ar-

tifact correction (Table I), however, resulted in distor-

tion of the cap. As shown in Figure 8, when not

corrected, rings were prominent. Rings diminished as

correction values increased and mostly disappeared at

the correction value of 12. Cap distortion became visi-

ble at correction value 6 and worsened at correction

value 12. Because cap distortion blurred the exterior

bounds, correction value 3 was more appropriate, and

correction values of 2–4 were used for image recon-

struction for all samples.

Figure 6

Effect of vial placement on RSF measurement: (a) Group 1 (randomly oriented vials) displays greater variation

compared with (b) Group 2 (vials with fixed orientation), when measured for RSF as seen in (c) %RSD plotted

for each vial.



3.4.2. Relationship of RSF Variability to Component

Positional Offsets: Scans were imported into a special-

ized software (ScanIP) that facilitates the generation

of 3D models and the separation of the models into

Figure 7

Micro-CT images generated from the same vial

under (a) enhanced scanning parameters (24-h scan

time) and (b) selected scanning parameters (4-h

scan time). Key features remain visible in both

scans.

Figure 8

Impact of ring artifact correction on reconstructed

slices. Applying aggressive ring artifact correction

results in cap distortion as seen in value 6 and most

notably in value 12. Because rings did not interfere

with analysis, a correction value of 3 was selected.

Figure 9

Top, bottom (Bot), and side views of a 3D model of

a scanned vial before and after segmentation. Seg-

mentation provides a means to analyze the individ-

ual features of interest (exterior bounds) on both

the cap and stopper.

Figure 10

Schematic for centroid offset distance analysis on two

sample scenarios. (a) Perfectly centered stopper and

cap display two concentric circles (top view); therefore,

no offset between centroids. (b) Offset stopper and cap

display two nonconcentric circles (top view); therefore,

a distance can be quantified between centroids.



separate components (i.e., the cap and the stopper).

The segmentation capability allowed individual features of

interest on the cap and the stopper to be analyzed sepa-

rately. Figure 9 shows a 3D model before and after segmen-

tation of a capped vial (button removed) from different

view angles. The exterior bounds of the stopper and the

cap, at approximately the midpoint of the stopper flange

thickness, could be fitted into two circles by the software

from the constructed image slice. Note that the selection of

the midpoint of the stopper flange thickness is to minimize

interferences from the stopper/vial and/or stopper/cap inter-

faces. Thus, component positional offsets within a vial

could be quantified by comparing the concentricity of the

two circles (Figure 10). When the stopper and the cap are

perfectly centered, the two circles are concentric (Figure

10a). If the stopper and the cap are slightly offset, the two

circles are nonconcentric (Figure 10b). Positional offsets

are quantified by the distance between the centroids of the

fit circles (see the offset illustration in Figure 10).

To verify that our selected scanning parameters were

accurate for analysis, a vial was scanned using the

selected scanning parameters and the enhanced scanning

parameters (discussed in 3.4.1). Their respective offsets

were calculated (Table IV). The fit circles’ centroids dif-

ferentiated by 18 lm, which is approximately the dimen-

sion of one voxel, suggesting that the acquisition,

reconstruction, and analysis for both scans were compa-

rable despite reduced image resolution.

The 10 selected vials were scanned. Their offset distan-

ces are summarized in Table V and plotted in Figure

11. The five vials with less RSF variability demon-

strated less offset distance compared with the five vials

with more RSF variability. The average offset distance

between these two vial groups is more than double and

statistically significant (p< 0.05). These data con-

firmed that component positional offsets are a major

contributor to intravial RSF variability.

TABLE IV

Offset Distance Analysis of Enhanced and Selected Scanning Parameters for the Same Vial

Feature of

Interest

Fit Circle Properties Distance between

Centroids (mm)Radius (mm) Centroid Coordinates (x, y)

Enhanced Cap 10.58 (10.67, 10.71)

Stopper 10.34 (10.83, 10.62)179

Selected Cap 10.56 (10.68, 10.79)

Stopper 10.32 (10.55, 10.88)161

Difference 18

TABLE V

Offset Distance Analysis for the Five Least and Most Variable Vials

Vial Number %RSD Distance between Centroids (mm)

Least variable 1 5.05 41

3 6.80 11

8 6.21 72

12 4.29 191

14 5.42 59

Average 5.61 75

Most variable 5 11.73 178

7 13.20 241

13 13.40 169

15 15.67 161

20 11.76 90

Average 13.15 170



4. Conclusions

The correlation of RSF to CCI established with serum

stoppers was consistent with that previously observed

with lyophilization stoppers (14). The superior CCI

performance of the lyophilization stopper under no or

low RSF lies in its stopper design, particularly in the

plug area (feature 6 in Table II). The time factor of

the RSF/CCI correlation was also assessed. The vis-

cous flow (relaxation) of the rubber stopper enabled

improvement of CCI over shelf life for some vials

with low RSF. This observation suggested that to estab-

lish the low RSF limit for a new CCS, RSF and CCI should

be tested soon after capping to ensure worst-case conditions

to be evaluated. While the RSF/CCI relationship was

established, selected groups of capped vials with

most and least RSF variability were evaluated for

positional offsets by a novel visualization methodol-

ogy, micro-CT. Although time consuming and costly,

micro-CT provided direct evidence that inherent tol-

erances associated with each container closure com-

ponent are inevitable and may cause positional

offsets. Thus, RSF measurement variability is a real-

ity, and a reasonable expectation should be set on the

application of RSF as a quantitative screening tool or

for GMP manufacturing use.

Acknowledgments

The authors would like to thank the teams at Micro

Photonics Inc. and Synopsys Simpleware Product

Group for their suggestions with micro-CT scanning,

reconstruction, and 3D analysis.

Conflict of Interest Declaration

The authors declare that they have no competing interests.

References

1. Morton, D. K.; Lordi, N. G.; Troutman, L. H.;

Ambrosio, T. J. Quantitative and Mechanistic

Measurements of Container/Closure Integrity.

Bubble, Liquid, and Microbial Leakage Tests. J.

Parenter. Sci. Technol. 1989, 43 (3), 104–108.

2. Lam, P.; Stern, A. Visualization Techniques for

Assessing Design Factors That Affect the Interaction

between Pharmaceutical Vials and Stoppers. PDA J.

Pharm. Sci. Technol. 2010, 64 (2), 182–187.

3. Brown, H.; Mahler, H. C.; Mellman, J.; Nieto, A.;

Wagner, D.; Schaar, M.; Mathaes, R.; Kossinna,

J.; Schmitting, F.; Dreher, S.; Roehl, H.; Hem-

minger, M.; Wuchner, K. Container Closure Integ-

rity Testing—Practical Aspects and Approaches in

the Pharmaceutical Industry. PDA J. Pharm. Sci.

Technol. 2017, 71 (2), 147–162.

4. Degrazio, F. L. Holistic Considerations in Opti-

mizing a Sterile Product Package to Ensure Con-

tainer Closure Integrity. PDA J. Pharm. Sci.

Technol. 2018, 72 (1), 15–34.

5. Ludwig, J. D.; Davis, C. W. Automated Method

for Determining Instron Residual Seal Force of

Glass Vial/Rubber Closure Systems. Part II. 13

mm Vials. PDA J. Pharm. Sci. Technol. 1995,

49 (5), 253–256.

6. Ludwig, J. D.; Nolan, P. D.; Davis, C. W. Auto-

mated Method for Determining Instron Residual

Seal Force of Glass Vial/Rubber Closure Systems.

J. Parenter. Sci. Technol. 1993, 47 (5), 211–253.

7. Morton, D. K.; Lordi, N. G. Residual Seal Force

Measurement of Parenteral Vials. II. Elastomer Evalu-

ation. J. Parenter. Sci. Technol. 1988, 42 (2), 57–61.

8. Morton, D. K.; Lordi, N. G. Residual Seal Force

Measurement of Parenteral Vials. I. Methodology.

J. Parenter. Sci. Technol. 1988, 42 (1), 23–29.

Figure 11

Box plot of offset distances for the five least and

most variable vials from Group 1 (random ori-

ented). Increased vial variability (%RSD) corre-

lates to greater offset distances.



9. Mathaes, R.; Mahler, H. C.; Buettiker, J. P.;

Roehl, H.; Lam, P.; Brown, H.; Luemkemann, J.;

Adler, M.; Huwyler, J.; Streubel, A.; Mohl, S. The

Pharmaceutical Vial Capping Process: Container

Closure Systems, Capping Equipment, Regulatory

Framework, and Seal Quality Tests. Eur. J.

Pharm. Biopharm. 2016, 99, 54–64.

10. Mathaes, R.; Mahler, H. C.; Roggo, Y.; Huwyler,

J.; Eder, J.; Fritsch, K.; Posset, T.; Mohl, S.;

Streubel, A. Influence of Different Container

Closure Systems and Capping Process Parame-

ters on Product Quality and Container Closure

Integrity (CCI) in GMP Drug Product Manufactur-

ing. PDA J. Pharm. Sci. Technol. 2016, 70 (2),

109–119.

11. Mathaes, R.; Mahler, H. C.; Roggo, Y.; Ovadia,

R.; Lam, P.; Stauch, O.; Vogt, M.; Roehl, H.;

Huwyler, J.; Mohl, S.; Streubel, A. Impact of Vial

Capping on Residual Seal Force and Container

Closure Integrity. PDA J. Pharm. Sci. Technol.

2016, 70 (1), 12–29.

12. Mathaes, R.; Mahler, H. C.; Vorgrimler, L.; Stein-

berg, H.; Dreher, S.; Roggo, Y.; Nieto, A.; Brown,

H.; Roehl, H.; Adler, M.; Luemkemann, J.; Huwy-

ler, J.; Lam, P.; Stauch, O.; Mohl, S.; Streubel, A.

The Pharmaceutical Capping Process—Correlation

between Residual Seal Force, Torque Moment,

and Flip-off Removal Force. PDA J. Pharm. Sci.

Technol. 2016, 70 (3), 218–229.

13. Zeng, Q.; Zhao, X. Time-Dependent Testing Eval-

uation and Modeling for Rubber Stopper Seal

Performance. PDA J. Pharm. Sci. Technol. 2018,

72 (2), 134–148.

14. Ovadia, R.; Streubel, A.; Webb-Vargas, Y.;

Ulland, L.; Luemkemann, J.; Rauch, K.; Eder, J.;

Lam, P.; Tegoulia, V.; Maa, Y. F. Quantifying the

Vial Capping Process: Residual Seal Force and

Container Closure Integrity. PDA J. Pharm. Sci.

Technol. 2019, 73 (1), 2–15.

15. Haeuser, C.; Goldbach, P.; Huwyler, J.; Friess, W.;

Allmendinger, A. Imaging Techniques to Charac-

terize Cake Appearance of Freeze-Dried Products.

J. Pharm. Sci. 2018, 107 (11), 2810–2822.

16. Hindelang, F.; Zurbach, R.; Roggo, Y. Micro

Computer Tomography for Medical Device and

Pharmaceutical Packaging Analysis. J. Pharm.

Biomed. Anal. 2015, 108, 38–48.

17. Nieto, A.; Roehl, H. Sealing Behaviour of Con-

tainer Closure Systems under Frozen Storage Con-

ditions: Nonlinear Finite Element Simulation of

Serum Rubber Stoppers. PDA J. Pharm. Sci.

Technol. 2018, 72 (4), 367–381.

18. Nieto, A.; Roehl, H.; Adler, M.; Mohl, S. Evalua-

tion of Container Closure System Integrity for

Storage of Frozen Drug Products: Impact of Cap-

ping Force and Transportation. PDA J. Pharm.

Sci. Technol. 2018, 72 (6), 544–552.

19. Anas, E. M.; Lee, S. Y.; Hasan, K. Classification

of Ring Artifacts for Their Effective Removal

Using Type Adaptive Correction Schemes. Com-

put. Biol. Med. 2011, 41 (6), 390–401.



Authorized User or for the use by or distribution to other Authorized Users·Make a reasonable number of photocopies of a printed article for the individual use of an·Print individual articles from the PDA Journal for the individual use of an Authorized User ·Assemble and distribute links that point to the PDA Journal·Download a single article for the individual use of an Authorized User·Search and view the content of the PDA Journal permitted to do the following:Technology (the PDA Journal) is a PDA Member in good standing. Authorized Users are An Authorized User of the electronic PDA Journal of Pharmaceutical Science and

copyright information or notice contained in the PDA Journal·Delete or remove in any form or format, including on a printed article or photocopy, anytext or graphics·Make any edits or derivative works with respect to any portion of the PDA Journal including any·Alter, modify, repackage or adapt any portion of the PDA Journaldistribution of materials in any form, or any substantially similar commercial purpose·Use or copy the PDA Journal for document delivery, fee-for-service use, or bulk reproduction orJournal or its content·Sell, re-sell, rent, lease, license, sublicense, assign or otherwise transfer the use of the PDAof the PDA Journal ·Use robots or intelligent agents to access, search and/or systematically download any portion·Create a searchable archive of any portion of the PDA JournalJournal·Transmit electronically, via e-mail or any other file transfer protocols, any portion of the PDAor in any form of online publications·Post articles from the PDA Journal on Web sites, either available on the Internet or an Intranet,than an Authorized User· Display or otherwise make any information from the PDA Journal available to anyone otherPDA Journal·Except as mentioned above, allow anyone other than an Authorized User to use or access the Authorized Users are not permitted to do the following:


Quantifying the Vial-Capping Process: Reexamination Using ... · 2.6.VialLandSealStudies Serum...

Documents

Transcript of Quantifying the Vial-Capping Process: Reexamination Using ... · 2.6.VialLandSealStudies Serum...