OR WAIT null SECS
The authors describe a validation master plan for closed-vial filling technology.
Continuous improvement in the development of technologies for aseptic filling of injectable products has led to innovative solutions. The main driver for innovation is that a small deviation from GMP may trigger product contamination leading to septic shock and even death for the patient. For example, analysis of an outbreak database has shown that, among 1537 patients contaminated by parenteral products, the mortality rate reached 15%. Of those contaminations, 20% were due to deviations during pharmaceutical drug manufacturing, showing that improvement of injectable processing operations is worth investigating (1).
The rate of contamination has been strongly reduced during the past 50 years thanks to major innovations targeting both the container and the filling technologies. Among the major changes observed during this period were:
Crystal closed-vial technology aims to further address the quality issue for the patient and also to simplify the aseptic filling process (2, 3). The concept of crystal closed-vial technology consists of three different steps:
Although the goal is still to fill a vial with the pharmaceutical drug under aseptic conditions, closed-vial technology allows for some differences. Three main advantages include:
Eliminating some process technologies requires the introduction of other solutions, such as:
Making such changes raises several questions from the pharmaceutical industry regarding their acceptance by regulatory authorities. To ensure that the technology is suitable for aseptic filing of injectables, a series of tests should be performed regarding container materials and characteristics, container manufacturing process, and filling process.
This article describes the major changes that have been introduced, the validation processes, and their results. These data have been used by GSK Biologicals to support the approval for Europe of Synflorix, a vaccine against pneumococcal infection, using the closed vial for a Type II variation of the previous approval of the same product in other containers. It is clear that product stability data are mandatory to obtain regulatory product approval, but providing a complete set of supporting data can help the applicant to set up a complete and consistent file for submission to the authorities.
As described above, several modifications have been introduced compared with the classical glass-vial technology. To ensure that these modifications meet regulatory requirements and expectations, they have been tested according to the US and European pharmacopeas and to the guidelines published by the International Conference on Harmonization (ICH), and compiled in the validation master plan.
The tests performed are listed in Table I. They can be split into four parts: the container–closure materials, the behavior of the container closure in various conditions, the specific filling operations, and the overall process.
Table I: Overview of the tests performed in the validation master plan.
One crucial aspect of the container–closure is that the container has to be treated by gamma-irradiation to ensure its overall sterility, and not only asepsis. Gamma-irradiation may have physico-chemical effects on the materials. Therefore, to be robust, validation must be performed using worst-case conditions regarding the irradiation dose. To that end, a dose-mapping experiment on the densest pallet (1-mL vials) showed that to ensure a minimum of 25 kGray at the coldest point, the hottest point can receive up to 46 kGray. Therefore, the decision was made to irradiate all the test material at minimum 50 kGray.
Different chapters of the pharmacopeia are applicable to closed-vial materials, as indicated in Table I. Of course, these tests must be conducted only on materials in direct contact with the pharmaceutical drug (i.e., the TPE stopper and the COC vial body).
The toxicity test described in USP <87> Biological Reactivity Tests, In Vitro has been conducted on both vial body and stopper materials. Extract of materials are subjected to the elution test in which L-929 cells are incubated for 48 hours with media used to extract materials. This test was repeated on three batches of each material and did not show any sign of reactivity.
The toxicity test described in USP <88> Biological Reactivity Tests, In Vivo is directly linked to the USP Class VI classification. It combines systemic injection, intracutaneous injection, and intramuscular implantation tests with extracted materials or strips of materials. Based on the results obtained, the material can be classified from Class I up to Class VI. According to the application, a certain level of classification is mandatory (e.g., USP Class VI is mandatory for permanent implantation such as prosthesis). For injectable containers, Class IV is acceptable. Nevertheless, as most of container materials meet the most stringent conditions (i.e., class VI), this one became a recommended standard. To perform a complete range of tests, the products must be extracted with four different solutions: 0.9% sodium chloride, 1:20 alcohol solution in sodium chloride, 0.9% polyethylene glycol 400, and vegetable oil. For both vial body and stopper materials and with all extraction solutions, the test was performed on three different batches and no toxicity was recorded, showing that USP Class VI requirements are met.
Other tests were performed regarding material physico-chemical properties. The US Pharmacopeia requires a series of tests on polymers in USP General Chapter <661>. These tests are limited to nonvolatile residues, residues on ignition, heavy metals, and buffer capacity. Some polymers such as polyethylene require specific tests but these are not applicable to COC. In the European Pharmacopoeia, chapter EP 3.1.3 Polyolefines, is dedicated to polyolefin. Although COC does not fall in that category, because the definition of polyolefin is limited to "polyethylene or co-polymer of ethylene with less than 25% of another organic compound," it was still decided to conduct that series of tests on COC. The vial body passed all these tests.
Concerning the tests to be conducted on the elastomeric closure, because USP <381> Elastomeric Closures for Injections contains the in vitro and in vivo tests described here as well as other tests harmonized with those described in EP 3.2.9 Rubber Closures for Containers, EP 3.2.9 for single-dose container was selected. This chapter describes mixing chemical tests on elastomeric closure extracts and mechanical tests such as fragmentation and penetrability on the whole stopper. The TPE stopper successfully passed all these tests for single-use container.
Among the expectations regarding behavior of a container closure, some of them come from pharmacopeias or ICH guidelines while others are based on glass-vial standards. The list of the relevant tests and expected behaviors are indicated in Table I. Nevertheless, before starting validation of the container closure, it is crucial to ensure that it is appropriate to use the vial after gamma-irradiation. Gamma irradiation generates significant ozone. Because the vial is closed, the ozone cannot exit from the vial, and therefore, its disappearance results mainly from its natural degradation in oxygen. To ensure that degradation is sufficient to return to acceptable conditions, remaining concentration was calculated on the basis of ozone half-life (6). As half-life in atmospheric conditions is less than an hour, the ozone level would return back to normal concentration within 10 hours. In practice, such delay is much shorter than the delivery time from the irradiation site to the filling site, so there is no concern with ozone content.
According to both the US and the European pharmacopeias, the particle content inside a small volume injection container should not exceed 6000 for particles > 10 µm and 600 for particles > 25 µm (if the light obscuration particle count method is used). The method recommends careful withdrawal of the water for injection (WFI) content of several vials (to exceed 25 mL of test solution) and to pool it into a clean container. Usually, the stopper is carefully withdrawn to perform the WFI collection. In case of the closed vial, the stopper withdrawal being not feasible, the collection is performed by means of a syringe equipped with a 23 G needle piercing the stopper (which means that the particles generated during piercing of the stopper for product collection are included in the results). Particles generated by the movement of the syringe plunger are neutralized by a blank value of syringe movement.
Table II: Particles generated by complete process (from vial manufacturing up to product collection) in 2 mL crystal vials.
The results from a typical test, summarized in Table II, show that the particle content is much lower than the acceptable limits. In addition, these data are roughly two times lower than the particle content in glass vials measured according to the same procedure. This test is performed routinely to monitor the level of particle presence in vial batches.
Container–closure integrity was assessed on vials that were pierced, filled, and laser resealed but not capped. The objective was to assess in one test the integrity at the junction between the stopper and the vial body and also the quality of the laser resealing. The dye ingress challenge test was used for validation and also for regular performance of batch release tests.
The selected dye test was based on EP 3.2.9 under the denomination "self-sealing test." The normal procedure consists of immersing vials in methylene blue solution and challenging them with a –27 kPa vacuum for 10 minutes, followed by a return to atmospheric pressure for 30 minutes. To be more challenging, three successive cycles were applied, each of them consisting of a vacuum of –30 kPa for 30 minutes, followed by an overpressure of +15 kPa for 30 minutes. The results showed the absence of dye ingress.
Microbiological tests have been conducted as well. Tests based on immersion in solution with Brevundimonas Diminuta and Proteus Mirabilis bacteria showed no contamination of vials filled with Tryptic Soy Broth and laser resealed.
It is crucial to ensure that the endotoxin contamination of vials is acceptable because no depyrogenization process takes place in the overall vial manufacturing and filling process. To ensure that the level of endotoxin is acceptable, both the raw material pellets and the vials are tested for endotoxin presence with an acceptable limit of 0.25 EU/mL. The method used was the Limulus Amebocyte Lysate (LAL) test using Endosafe equipment from Charles River. All batches tested until now showed low level of endotoxin on raw materials (1 g extracted with 2 mL of WFI, 0.05–0.10 EU/mL for 5 batches of raw materials) and the level inside the vial was always below the detection limit (<0.05 EU/mL UI for all vial batches). If endotoxin levels appears to be systematically low thanks to a robust process to avoid contamination, the level of control may be reduced in the future following a risk analysis.
Because the vials are sterilized by gamma-irradiation, the bioburden should be assessed to ensure that the irradiation dose is in line with ISO-11137 requirements. Because both vial body and stopper are molded at a temperature in the range of 200 °C and directly assembled in ISO5, the bioburden should be very low. Currently, all the samples tested (10 vials from each tested batch) showed absence of bioburden. This test should be performed either each batch or every three months according to ISO-11137 guidelines.
Compared with glass, polymer materials have a significantly higher permeability to gas. COC material, used for the vial body, has been selected on basis of its well-recognized gas barrier properties. Nevertheless, gas permeability has to be assessed to ensure that products will not be subject to significant changes. In particular, water vapor transmission rate (WVTR) and oxygen exchange must be carefully checked. WVTR can lead to loss of water from liquid product, resulting in change of product concentration. Oxygen ingress can lead to oxidation of oxygen-sensitive products.
Table III: Water vapor transmission rate from 2-mL vials filled with 1.2 mL water-for-injection. RH is relative humidity.
To measure WVTR, 2-mL vials filled with 1.2 mL of WFI were stored in various ICH conditions. According to the ICH Q1A (R2) guideline, semipermeable containers should be stored in the following conditions to meet the class III and IV conditions (i.e., hot conditions): –20 °C for frozen products, 2–8 °C (normal conditions) and 25 °C ± 2 °C / 60% ± 5% relative humidity (RH, accelerated conditions) for refrigerated products, 3 °C ± 2 °C / 35% ± 5% RH (normal conditions) and 40 °C ± 2 °C / not more than 25% RH (accelerated conditions) for products kept at room temperature. According to ICH, a loss of 5% water within a period of three months in accelerated conditions is the acceptable limit. If that limit is exceeded, investigations should be conducted to assess the effect on the product.
Table III illustrates the loss of water in various storage conditions. In both accelerated conditions, water loss was significantly below the limit. Similar experiments are on-going with other vial formats and identical conclusions can be drawn. This supports the initial expectation that COC has excellent barrier properties for water vapor.
It is less obvious how to measure oxygen permeation due to ambient oxygen. Therefore, a specific test was set up to characterize the vial behavior regarding permeation to oxygen. The vial was packed with a bag of FeO crystals, validated regarding absence of loss of crystals outside the pack, inside a sealed aluminum pouch to serve as an oxygen scavenger. Such an assembly will first deplete the oxygen around the vial, bringing the oxygen concentration inside the pouch in the range of 0.1% within 24 h. In a second phase, the oxygen will leak out of the vial because of the concentration differential. When exiting the vial, the oxygen is captured by the scavenger, keeping the external concentration permanently low. Figure 1 shows the different kinetics observed according to storage temperature, leading to residual oxygen concentration in the vial of less than 3% in two months.
Figure 1: Speed of oxygen depletion in 2-mL closed vials when incubated with an oxygen scavenger in aluminum pouch. Oxygen concentration was measured using a Microx TX3 fiber optic oxygen transmitter (Presens), connected to a needle-type oxygen microsensor. Vials were incubated at 25 Â°C (blue line) and 40 Â°C (red line).
A useful application of that permeation to oxygen is the opportunity to prepare oxygen-depleted vials in advance after ensuring sufficient preincubation time. When filling oxygen-depleted vials by piercing with the noncoring needle through the stopper, no significant oxygen increase was observed in the vial, maintaining the initial oxygen concentration without nitrogen flush. For example, a test made with 2 mL oxygen depleted vials (pre-incubated for four months) showed that before and after filling, the concentration remained around 0.5% oxygen. After filling, the vials were kept under aluminum pouch protection with an oxygen scavenger for six months. The oxygen concentration inside the vials continued to decrease down to 0.1%. Although it requires an additional secondary packaging, closed vials are able to decrease the level of oxygen to significantly lower levels than classical procedures applied to glass vials, showing significant benefit for oxygen-sensitive products.
To define the shelf-life of vials, irradiated vials were stored for up to five years. On a regular basis, vials were tested for appearance, container closure, particle content, sterility, and laser reseal ability. Until now, four years of shelf-life have been validated for the 2-mL vials.
Extractables studies are becoming more and more mandatory. The purpose of the extractables studies is to identify all compounds that could leach out of both the vial body and the stopper, while leachable studies aim to identify all compounds that actually leach to the solution. Therefore, the detection of an extractable does not mean that this compound will be present as a leachable when long-term studies are conducted with pharmaceutical drugs, but it is worth specifically tracking such a chemical compound when the possibility of it being present exists. To perform appropriate extractable studies, four factors must be taken into account:
For each of these key success factors, a method has been set up to ensure optimal extraction.
Based on extractable results, a list of 13 potential compounds categorized as potential leachables was established and these compounds were followed during compatibility studies.
Beside drug stability studies, the behavior of the vial itself must be assessed because its performance may vary according to the solution filled inside. Some products can be aggressive to the materials used for the container, leading to significant leachables, while others can degrade components material. Before final testing with pharmaceutical drugs, preliminary supporting data can be collected with some of the classical solvents/excipients. The characteristics assessed during long-term compatibility studies are visual aspect, maintenance of closure integrity, particle generation, and leachables.
The first three tests were conducted in the presence of WFI but additional simulants were selected for leachable studies. These simulants were WFI, 10% ethanol, phosphate buffered saline, 0.9% NaCl, and 0.5% 2-phenoxyethanol. Vials of 2 mL were filled with 1.2 mL of simulant and stored 50% right-side-up and 50% upside-down. The last results corresponded to two years of incubation at 30°C ± 2°C / 35% ± 5% relative humidity and showed that the visual aspect of the vial was not altered, the closure integrity was maintained, and the particle profile did not change significantly. Regarding the leachables, a one-year assessment was conducted to notice possible differences. Because no major differences were seen after one year, two of them, WFI and 10% ethanol were selected for long-term studies (five years). The leachable results showed the presence of two compounds, acetic acid and formic acid, exceeding 5 ppm. Low leachable quantities of t-butanol and acetone were detected in concentrations exceeding 0.2 ppm and all other products, including nine of the extractables recommended for follow-up, were either not detectable or at trace level. It is important to note that these four most abundant compounds in solution belong to the Class III solvent category according to the ICH Q3C guidelines. In these guidelines, the acceptable daily administration dose for Class III solvents is fixed at 50 mg, equivalent to more than 40,000 ppm in a vial filled with 1.2 mL. In addition, the full leachable profile was reviewed by a toxicology expert who did not report any toxicity concern.
Low temperature storage
Many of the new pharmaceutical drugs belong to the biological class. These products are often sensitive to heat and show lack of stability even at 2–8 ° C (for example, recombinant viruses and cell therapies). One option is to store them frozen at –20 °C, at –70 to –80 °C or even in liquid nitrogen (in vapor or submerged inside, i.e., at –196°C).
Some issues are reported with glass vials when stored at very low temperature, such as small cracks —ranging up to complete breakage—and loss of closure integrity caused by differential retraction of the stopper regarding the glass vial neck. Loss of integrity can lead to contamination risk, loss of sublimated water, or severe acidification when stored in dry ice.
To assess the suitability of the vial for low temperature storage, WFI-filled vials were stored as long as one year at –20°C, –80°C, or inside liquid nitrogen. The outcome is that none have shown any visual defect, loss of closure integrity, or loss of water.
To confirm the vial integrity at very low temperature, vials were filled with 100% oxygen and laser-resealed before being immersed in liquid nitrogen. Some unsealed vials were equipped with a small catheter in the stopper as positive samples. After 11 days of storage in these conditions, the oxygen content remained at 100% in resealed vials, whereas it decreased to about 60% in positive vials. This result showed that nitrogen did not enter the resealed vials, thus demonstrating the maintenance of the tightness of the closure.
These tests demonstrate that the closed vial, because of its resistant COC and effective closure assembly, is suitable for long-term storage in extreme conditions of low temperature.
Resistance to breakage
COC is known to be resistant to damage, and resistance to breakage was assessed in two different conditions: by drop test, to assess risk of cracks due to shock, and by freeze-thawing to assess resistance to the expansion of frozen liquids.
To quantify resistance to shock and to compare it with that of glass vials, a drop test procedure from different fixed heights to a concrete floor was set up. The percentage of intact and nonleaking vials was recorded for each height. The results showed that for the two tested sizes (1-mL and 10-mL vials), the closed vial is significantly more resistant than the equivalent glass vial (2R and 10R vials). For example, for the 10 mL vial size, 50% of glass vials are broken when falling from table height (75 cm) whereas closed vials can stand up to above 2 meters before seeing the first damage (see Figure 2).
Figure 2: Damage to 10-mL closed vials (blue line) and 10R glass vials (red line) when dropped from different heights. Vials were dropped once from each height, starting from the lowest, until damaged or broken.
To assess resistance to product expansion, a freezing-thawing test at –80 °C was conducted with vials filled with 15% mannitol, which is known to expand significantly during freeze-thawing. This test was performed according to a test reported by Jiang et al. who recorded glass vial breakage when frozen with this solution inside (7). Jian et al. observed that glass vial breakage depends on the filled volume and on the vial size. Jian et al. recorded 100% breakage for 20-mL glass vials and 50–60% breakage with 5- and 10-mL glass vials when filled at half of nominal volume. When this test has been performed according to the same procedure, no broken closed vial was observed with 5-, 10- and 20-mL size. The test was also performed with closed vial filled at full nominal volume and again, no breakage was reported.
This test showed that not only is COC resistant to shocks but that it is also able to tolerate significant product dilatation without damage.
The main changes in equipment compared with classical filling technologies are the needle piercing of the stopper and the laser resealing of the piercing trace.
Regarding the validation of the needle, the three main aspects assessed were the venting of the overpressure created inside the vial during filling, the level of particle generation during piercing, and the ability of the needle to support thousands of piercing.
Venting of the overpressure was validated by using a manometer connected to the inside of the vial. It showed that the pressure increased during filling in the range of 0.1–0.2 bar according to filling speed and vial size, then returned back to normal atmospheric pressure within 0.2 to 0.4 seconds. This measure correlates well with the observation that a short delay is necessary before needle withdrawal to avoid drop generation due to filling tube expansion, especially if silicone tubing is used.
Results of the particulate analysis showed that the needle generates very limited amount of particles during the piercing process. This result was obtained because of the noncoring pencil point design of the needle.
Moreover, the evolution of the particle generation inside the vial along a 5000-vial filling campaign using the same needle, did not show any increasing trend that could have been induced by the needle wear. Being observed with a microscope, the needle did not show any difference in shape, cutting edge or surface quality.
The laser needs to be validated on several aspects. First, the absorption coefficient of the stopper material must be characterized; second, the impact on material should be assessed; third, the effectiveness of the resealing should be validated and last, the absence of effect of the laser on the product should be verified.
The TPE characteristics should ensure absorption of the laser beam energy by the stopper in such a way that the needle trace is resealed to a sufficient depth without burning the surface of the stopper. Simultaneously, almost all the laser beam energy must be absorbed in the stopper thickness to avoid the laser beam passing through the stopper and hitting the filled product. The right absorption rate is obtained because of a colorant master batch mixed with the stopper material. The range of colorant percentage in stopper composition is precisely defined to obtain absorption of at least 97% of the laser energy over the thickness of the stopper, according to a Beer-Lambert curve.
The potential effect of the laser shot on the stopper material was assessed by near-infrared spectrometry of the stopper surface before and after a laser shot. The profiles are identical between the two records, showing that there is no detectable change of stopper composition.
The laser power to be selected for an effective laser resealing was validated with physical challenge tests on the closure integrity of the resealed stoppers: dye test on vials resealed at different laser powers and bubble test looking for leakage from over-pressurized vials. It was shown that the stoppers that were not resealed only exhibit a small percentage of failure at the dye test while stoppers resealed with enough energy do not fail. Another way to assess the effectiveness of the resealing is to cut the resealed stopper perpendicularly to the trace. The resealing depth ranged from 0.4 mm to 1.0 mm when using a 6 mm diameter 12W laser beam. To assess the resistance of the weld, overpressure was produced inside the vials through their bottom side and leakage was assessed by bubble generation at the piercing trace. Resealed vials started leaking from 140 kPa to 200 kPa whereas a range from 30 kPa to 150 kPa was observed for the nonresealed vials.
To ensure that the product was securely stored when laser resealed, the effect of a laser shot on the temperature was measured at different location of a 2-mL vial and results are illustrated in Table IV.
Table IV: Temperature change recorded at various position of the vial when a 4 mm laser shot is applied.
The results, achieved using a 4 mm diameter laser and reproduced with a 6 mm diameter laser, showed that the product was fully secured inside the vial because no temperature change could be detected following laser resealing of the stopper.
To challenge the new concept, media fills had to be run. The first media fills were performed with equipment in an assembly workshop. These conditions were challenging for several reasons including:
Despite these adverse conditions, none of the 26,313 TSB filled vials were contaminated during media fill. Contamination was assessed after vials were incubated for 14 d, during 7 of which they were stored upside-down. Since installation of that equipment in an ISO8 clean room, more than 88,000 additional vials have been filled with TSB without any contamination.
Developing a new technology for filling injectable products requires an in-depth validation plan to ensure that such technology meets all regulatory requirements and provides the necessary features to protect and conserve the product. This article has shown that the closed-vial technology satisfies all requirements according to regulatory authorities. This article describes a useful structure of investigation to follow to prove suitability for drug approval when developing a new container (or of a new part of container).
These data have been well received by the European authorities to support approval of Synflorix, a vaccine from GSK Biologicals against pneumococcal infections that has been approved in all European countries.
Aseptic Technologies receives grants from the Region Wallonne and from the Agence Wallone à l'Exportation (AWEX). Technology has been licensed by Medical Instill Technologies.
Benoît Verjans, PhD*, is chief commercial officer, Anne Glibert is in quality control, and Patrick Baleriaux is chief executive officer, all at Aseptic Technologies, Gembloux, Belgium. *To whom correspondence should be addressed, Benoit.firstname.lastname@example.org.
1. R.P. Vonberg and P. Gastmeier P, J. Hosp. Infect. 65, 15–23 (2007).
2. B. Verjans, J. Thilly, and C. Vandecasserie, C., Aseptic Processing, supplement to Pharm. Tech. 29, S24–S29 (2005).
3. J. Thilly, D. Conrad, and C. Vandecasserie, Pharmaceutical Engineering 26, 66–74 (2006).
4. B. Verjans and C. Reed, BioPharm Int. 25 (3), 46–58 (2012).
5. B. Verjans, Pharm. Tech. Eur. 24 (6) 49–55 (2012).
6. "Ozone," http://en.wikipedia.org/wiki/ozone, accessed Aug. 2012.
7. Jiang et al., PDA J. Pharm. Sc. Technol. 61, 441–451 (2007).
8. B. Verjans et al., Pharm. Tech. 31 (4), 184–195 (2007).