We discussed previously how Container Closure Integrity is critical for the sterility and the safety of the product. Because of this, there has been a constant push during last decades to develop new technologies and control systems able to check integrity, find and reject containers having dangerous leaks.
The standard regulation commonly used for reference is USP
<1207> even though it has not yet been updated and is still referring to
older estimates of MALL (Maximum Allowed Leakage Level).
In contrast with visual inspection for visible defects, in
CCIT we have to finds leakages which can cause external dirt or bacteria to
enter the container and compromise the product sterility. This involves more
difficulties than visual inspection because:
The leak can be in a non-visible
position, for example under the aluminum sealing of a vial
The leak can happen or become dangerous
only after the inspection and delivery of the product, for example because of
shocks during transportation or storage
The effect of a leak is accumulated over the whole shelf
lifetime of the product, therefore a small leak can be too small to be detected
during inspection, but can over the following weeks and months let enough
non-sterile air to enter.
Therefore careful visual inspection after manufacturing is
not enough to guarantee CCI. It must
be accompanied by a careful assessment
of the manufacturing process and all components involved (QA) and
make use of non-visual inspection in order to detect every possible leak.
Figure 1: Defects found divided by category
Inline and Offline Testing
Similarly with what we saw before for particles, even leak
testing can be executed either in sampled, laboratory, controlled, slow and
destructive way or in 100%, in-production, quick and non-destructive way.
The off-line, laboratory control can be much more accurate but is obviously not suitable for 100% production control which is already mandatory by main pharmacopeias for all fusion-sealed containers and will most likely be required for any other container in the near future.
Name |
Inline/Offline |
Speed |
Accuracy |
Cost per container inspection |
Blue-dye |
Lab
only |
Very
low |
High |
Medium
to high |
Microbial
Ingress |
Lab
only |
Very
low |
High |
Medium
to high |
Helium / tracing gas mass spectrometry |
Lab
only |
Very
low |
Very
High |
Very
High |
Vacuum/pressure decay |
Lab
and Inline
100% |
Low
in the lab High
inline |
Depends
on testing speed. Medium |
Medium
to low |
Laser headspace analysis (HGA) |
Inline
100% |
High |
Can
be high but requires modified headspace and resting time |
Low |
High Voltage (HVLD) |
Inline
100% |
High |
Can
be high but depends on liquid and position |
Low |
Other systems (sealing force, mass
loss, compression plates) |
Inline |
Medium
to high |
Low
to medium |
Medium |
The main technologies available today for leak detection
can be split in two main categories according to their in-line or off-line
suitability and are listed in the table below together with their main features.
Even without entering more in details, it is clear from
table above that no CCIT technology can be considered "the best". The
results and performances of each system depend heavily on the features of
product and manufacturing context.
We can highlight the main technologies for leak detection and their features and application in today's manufacturing of parenteral drugs.
The most widely used technology is HVLD (High Voltage
Leak Detection)
This is based on the application of a high voltage field
across different areas of the container between inside and outside. If a spark
and consequent flow of current is detected, it means that there is a crack and
the container is rejected.
This technology is well established, reliable and compatible with high production speeds. But it can work only with liquids having a minimum level of conductivity. Pure water and freeze dried products cannot then be inspected in this way.
Other limitations related to this technology are:
·
the
result is basically on/off, without any indication about the position or the
size of detected crack
·
the
application of a high voltage field can degrade the product or create ozone
· the sensitivity of the system changes in different areas of the container's
Another technology used inline is PDLD (Pressure Decay
Leak Detection)
In this case the container is enclosed into a sealed
chamber and then a higher or lower pressure is applied and measured with high
precision during inspection time. If there is a leak, the pressure into the
sealed chamber will change because of flow of air into or from the container.
Compared to HVLD this system can be used with any type of
hard container and solid/liquid product inside.
The main drawbacks are:
A higher mechanical complication and
cost
The detection of leaks is also
influenced on the position because the liquid inside can easily clog the crack
Finally, the most modern technology is HGA (Headspace Gas
Analysis)
If the adsorption of laser by oxygen or moisture in the
headspace is too high, it means that the headspace vacuum or inert gas have
been altered by outside air and the container is rejected.
The main drawbacks of this technology are:
Can work only with modified headspace,
either with lower pressure or filled with inert gas such as Nitrogen
In order to have a good sensitivity, the products must undergo a holding/waiting time to allow the headspace to be altered by the leakage. To reach a high sensitivity, the holding time can become very long.
Validation of Leak Detection Systems
Since CCIT cannot be performed effectively by visual inspection, the 100% inspection must be performed by automatic systems, most of times one of the three described before (HVLD, PDLD, HGA).
To validate this automatic system the main obligation given by pharmacopeia (US FDA) is that “the container-closure system to maintain the integrity of its microbial barrier, and, hence, the sterility of a drug product throughout its shelf life”.
The first step is therefore to correlate the leak size, or better the flow-rate, with the sterility of the drug over a long enough time span by using microbial or blue dye ingress systems. This will provide our MALL (Maximum Allowable Leak Level) value.
Then we need to prepare several samples with different and known leak levels. This can be made by different systems: laser drilled pin holes in glass or metal, micro-capillaries, micro-pipettes, copper wire.
The problem here is that while relatively large leakages, and equivalent pin-holes larger than 20/30μm, can be made quite easily and precisely, for smaller and smaller sizes the flow rate of leakages is no more stable and controllable because it depends heavily on material's surface, depth and roughness of the pin hole / capillary / pipette.
Figure 2: 5μm laser drilled hole
In other words, if we prepare and test 10 vials with a single nominal 5μm pin hole laser drilled into glass and we compare their leak flow rate with a lab precise measuring system like Helium spectrometry, we will find a large difference between them and the expected theoretical calculated flow rate.
Figure
3: Blue dye ingress test
This issue is making more and more difficult to have samples with precise and reliable leaks to be used for testing and validating the inline inspection system.
From these considerations we can understand how CCI testing is critical for safety but also still very challenging in technological and validation aspects. The control systems involved are complicated and require long time and high cost to provide good and reliable results.
Figure 4: Helium spectrometry layout
Once more as we already saw in visual
inspection, the knowledge of specific features of products and manufacturing
process and the best cooperation between manufacturer and inspection system
supplier is the best way to face and solve all problems and obtain the best
results.