Abstract
The pharmacopeia method for measuring the chemical durability of parenteral glass packaging is the hydrolytic resistance test in which the vial is filled to 90% of its brimful volume as described, for example, in USP <660>. However, an increasing number of innovative drugs are filled significantly below the nominal volume of the vial. As a consequence, the determined hydrolytic resistance is not representative of the concentrations of leached “glass” elements for low fill volumes. This is attributable to two main factors: Firstly, an increasing ratio of the wetted surface to volume and secondly an increased leaching tendency typically observed with borosilicate glass of the wall near bottom area, especially when standard manufacturing technology is applied.
The extent of both contributing effects has been analyzed by determining the amounts of the representative leached “glass” elements, boron, sodium, and silicon, after vial storage for 24 weeks at 40°C with different fill volumes (0.5, 1.0, and 2.0 mL). The vials which have been investigated in this study have a nominal fill volume of 2 mL, were made from Type I class B borosilicate glass (Fiolax®) and from aluminosilicate glass and were filled with either purified water or a 15% KCl solution.
The standard conversion process for tubing into vials was used for Fiolax vials (standard quality vials) and for aluminosilicate vials. In addition, an optimized conversion process (delamination controlled technology) was used to create low-fill quality Fiolax vials. The vial quality obtained from the two different converting technologies greatly influenced the concentrations of leached “glass” elements measured, especially when low fill volumes were used.
LAY ABSTRACT: Borosilicate glass containers, because of their chemical inertness, excellent barrier properties, high transparency, and mechanical stability, have been successfully used for decades to package parenteral drug formulations. Nevertheless, Type I glass can be altered over a period of time when in contact with the drug formulation. The result of this interaction is even more pronounced for some new innovative drugs that are delivered to the patient in small dosages significantly below the nominal storage capacity of the glass vials. When the fill volume of the vials is reduced, the contribution of the bottom area to the wetted surface increases, resulting in a higher surface-to-volume ratio. Therefore, the concentrations of leached elements will be increased and this can cause problems for sensitive medical products. This effect is not usually observed with the standard test procedures described in the pharmacopeia because the vials are filled with a high volume to 90% brimful capacity (e.g., as described in USP <660>). In this study, the leachable behaviors of vials made of borosilicate and aluminosilicate glass were evaluated by using medium and low fill volumes with storage for 24 weeks at 40°C. The standard conversion process to manufacture a vial from glass tubing introduces volatile “glass” elements into the vial wall near the bottom area. This mechanism has been described and supported by time-of-flight secondary ion mass spectrometry (TOF-SIMS) measurements of the inner vial surface as reported by Rupertus et al. The diffusion mechanism of volatile components will increase the leaching propensity of the vial, especially for low fill volumes. However, innovative manufacturing techniques are able to avoid the diffusion of volatile elements into the wall near the bottom area. This is achieved by a specific process setup in combination with a suitable monitoring test during vial production, which gives a quantitative measure of the leaching tendency of the wall near the bottom area. Borosilicate glass vials manufactured with this setup (low-fill quality vials) showed a drastic reduction in leachables, especially with low fill volumes. Vials composed of a boron-free glass showed no advantages in terms of leaching behavior when compared with borosilicate glass vials in general.
Introduction
For many years, containers made from borosilicate glass tubing have proven their superior suitability for use as primary packaging for parenteral products. When filled with aqueous solutions with a near neutral pH, they are attacked very slowly. Nevertheless, leaching of the main glass components, such as sodium, boron, aluminum, and silicon, may occur over time depending on the drug formulation used for filling and on the storage conditions. Depending on their concentrations, these leachables could be the reason for a shift in pH, formation of particles or precipitates, generation of compound or adsorption layers on the glass interior surface, or initiation or acceleration of drug degradation (1, 2).
Two factors determine the amount of glass components that may be found in the product after a certain storage period. The first factor is the dissolution propensity of the container inner surface that results from the interaction between the glass surface and the filling medium. The majority of relevant studies rely on testing protocols that use an aqueous extraction medium and that monitor the leached “glass” elements and/or the shift in pH values (3). In addition to the glass composition, the leachables profile can also be influenced by potential surface treatments, the container format, and the forming process of the container. During the process of converting the tubes into vials, the bottom and the heel zone are intensely heated by gas burners causing some of the most volatile glass components to evaporate from the near-surface region. A relatively nonuniform composition of the glass surface is established depending on the particular parameters of the conversion process. Recently published studies revealed compositional variations in different areas within the topmost surface as determined by X-ray photoelectron spectroscopy (4) and within the near-surface layer as determined by SIMS depth profiling (5). Surface areas of vials with an altered glass composition will in general show a heightened leaching behavior or an increased delamination propensity compared to surface areas of unconverted tubular glass (6).
The second factor is the ratio of the container inner surface that is wetted by the product to the fill volume. This ratio will greatly increase with decreasing fill volume. The wetted surface comprises the altered bottom surface region and the unaltered cylindrical part of the container wall that has not been directly heated during the hot forming process. The bottom and the cylindrical parts next to the bottom will always be wetted, even if the fill volume is very low. As the fill volume increases, a larger portion of the unaltered cylindrical wall will be wetted in addition. For this reason, products that fill only a low volume of the container and are stored under normal conditions (bottom down) will be exposed to chemically altered glass surfaces to a much greater extent than products that fill a nominal container volume. Therefore, products that fill lower container volumes than nominal may suffer from leaching more severely because of geometric reasons.
Kucko et al. (7) have investigated this effect in vials filled with water for injection by determining the pH shift. Their study showed an obvious trend for significantly increasing pH with decreasing fill volume. Eight different types of vials, mainly of Type I class A and Type I class B glass, from four suppliers were investigated. All vials were autoclaved at 121°C for 1 h. On the basis of the resulting pH value measurements, the authors state that their “pH results…provide an initial indication that the heel zone of converted glass vials is more susceptible to aqueous corrosion.”
A wide variety of therapeutic proteins as for example monoclonal antibodies are filled with volumes below 2.0 mL. In particular for vaccines, two-thirds have fill volumes lower than the container nominal volume. Therefore, leaching of glass components in the case of low fill volumes has received increasing attention.
Materials and Methods
Materials
Three different vial types with a nominal fill volume of 2 mL were produced on state-of-the-art converting lines. Two vial types were made from borosilicate glass tubing whereas the third vial type was made from aluminosilicate glass tubing. The inner diameter of all of the glass tubes was 14.0 ± 0.2 mm (2R vial in the so-called “ISO-format”). Table I summarizes the main oxide components of the two glass types; the aluminosilicate glass does not contain boron oxide and the borosilicate glass does not contain magnesium oxide.
Main Oxide Components of the Glasses Used for the Vials
The first vial type (low-fill quality vial) was formed from Type I B borosilicate glass tubing (Fiolax) by using converting parameters based on the delamination-controlled conversion process (4). This optimized converting procedure ensures that the durability of the entire vial wall, including the critical wall near the bottom area, is nearly constant. The production control of these vials is performed with the so-called SCHOTT Quicktest (4, 6) with validated threshold values.
The second vial type (standard quality vial) was formed from Fiolax tubing by using a converting setup that ensured compliance with regulatory requirements, such as USP, Ph. Eur., and JP, but that exceeded the limits defined by the Quicktest. The Quicktest can be seen as a way to determine the increased vulnerability and the leaching tendency of the wall near the bottom area.
The third vial type (aluminosilicate vial) was formed from aluminosilicate glass tubing. The converting parameters were adjusted because of the different physical properties of the glass (e.g., transformation temperatures, viscosity profiles) but the same hot forming machines were still used. Neither the aluminosilicate glass tubes nor the vials were commercial products; they were manufactured in development projects with production equipment.
Prior to filling, storage, and analysis, the vials were labeled with a waterproof pen and cleaned in two steps: (1) fill and empty three times with tap water, and (2) fill and empty three times with purified water. A pipette was then used to fill the vials with two unbuffered solutions. All three vial types were filled with water and the borosilicate vials were also filled with a salt solution (KCl). Both solutions are described in more detail below:
Purified water R with a conductivity <5 µS/cm at 25°C.
Potassium chloride 15 wt % solution (15 g KCl dissolved in 100 mL purified water).
Both solutions were used at three different fill volumes (0.5 mL, 1 mL, and 2 mL) in the different vials, resulting in a total of 15 different samples types (see Table II). After the filled vials were capped with Westar RS B2-40 Fluro-Tec stopper manufactured by West Pharmaceutical Services, the vials were stored upright in an oven at 40°C for 24 weeks, which is close to the time period instructions given in the ICH Q1A (R2) guideline for accelerated studies (40°C for 24 weeks).
Overview of the Leachable Study Performed
The two filling media—water and KCl solution—were chosen because of their different ionic strengths but more or less similar neutral pH values. The core of the study is the comparative investigation of the behavior of two quality types of vials made of one of the most commercially available borosilicate glasses (Fiolax). To verify the influence of the glass type on the general leachable behavior, vials made of aluminosilicate glass filled with water were also analyzed, keeping in mind that these vials are at an "experimental stage" with respect to their glass composition and the converting process.
The data set was generated by combining the results of two successive studies. The first study (SCHOTT code PS-16-00,325) was started in 2016 and used standard quality and low-fill quality Fiolax vials filled with purified water and 15 wt % KCl solution. Based on the results obtained after 24 weeks, a second study (SCHOTT code PS-17-00,171) was started using aluminosilicate vials filled with purified water in order to generate a data set that was not influenced by borate evaporation.
The concentrations of the typical “glass” elements within the solutions as prepared and without storage (blank values) were analyzed using high-resolution inductively coupled plasma mass spectroscopy (HR-ICP-MS) by using a Thermo Scientific™ ELEMENT™ 2 tool for the purified water solution and using inductively coupled plasma optical emission spectroscopy (ICP-OES) with an Agilent ICP-OES 725 series instrument for the KCl solution. Only the amount of sodium in the 15% KCl blank solution had a concentration above the limit of quantification (LoQ) as listed in Table III. Ten vials were filled and stored for sample types with 1 or 2 mL fill volume, whereas 20 vials were filled and stored for samples with the 0.5 mL fill volume. Using this setup a total fill volume of at least 10 mL was available for each sample type.
ICP Results for Blank Solutions
Methods
Two different ICP methods were used to determine the concentrations of the typical “glass” elements in the solutions: HR-ICP-MS for purified water and, for reasons of accuracy, ICP-OES for 15% KCl solution. The data summarized in Tables IV and V are calculated as average from a double determination. The difference between the two single values was always within the measurement uncertainty (see Table III). The analytical methods for the purified water and the KCl solution were verified by a base validation.
Concentrations of Leached Elements with Purified Water Filling after 24 Weeks of Storage at 40°C
Concentration of Leached Elements with 15 wt % KCl Solution Filling after 24 Weeks Storage at 40°C.
After 24 weeks of storage, two pools for each condition of at least 5 mL (double determination) were created and diluted in 2% HNO3 solution before analysis. The dilution factors varied between 1:5 and 1:40 with respect to the concentration range of the samples. For quantitative evaluation, single element standards from Alfa Aesar were chosen. The mean recovery rate was found to be between 95% and 105%.
An assessment of the data based on geometric considerations helps to explain how the glass type and the converting process influence the behavior of the “glass” element concentrations when the fill volume is reduced. Neglecting any meniscus effects at the filling line, a filling height of ∼13 mm above the interior surface of the bottom is calculated for the 2 mL fill volume (inner diameter of the tube 14 mm). This leads to a wetted contact surface area of ∼7.2 cm2 for a 2 mL fill volume, whereas a wetted contact surface area of ∼2.9 cm2 is derived for a 0.5 mL fill volume. A detailed summary of the geometric parameters is given in Table VI. The wetted area to fill volume ratio increases when lowering the fill volume as the contribution of the bottom area is increasing.
All analyses were done with validated methods and qualified equipment within a laboratory that is accredited by the German accreditation agency DAkkS.
Results and Discussion
Two main contributing effects have an impact on the concentrations of the leached elements: First, the leaching propensity of the wetted area at the interior surface of the container in contact with the liquid as an important material-related factor, and second, the ratio of the wetted area to the fill volume as a significant geometric factor that leads to a higher concentrations of “glass” elements in the filling medium for low fill volumes, even if the leaching propensity of the surface remains the same. To understand the leachables profiles derived for different fill volumes, this geometric impact needs to be considered separately. A summary of the relevant parameters is given in Table VI for the various fill volumes with the vials (2 mL nominal filling with an inner diameter of 14 mm) used in this study. Similar data for a 3.6 mL fill volume are also included because this corresponds to the 90% brimful volume used for testing according to USP <660> (8).
Geometric Parameters of a 2 R ISO Vial
Assuming a constant leachable propensity, the geometric factor will result in a 1.63-fold increase in the concentration of leachables from the glass for 0.5 mL filling compared to 2 mL (wetted area to fill volume ratio of 5.9 cm−1 vs 3.6 cm−1).
Concentrations exceeding this expected “geometric” increase of 1.63-fold are an indication of regions with higher leachable propensities, which also means an indication of a reduced chemical durability. For a tube with infinite length and an inner diameter of 14 mm, the wetted area-to-volume ratio will be 2.9 cm−1.
The concentrations of sodium, silicon, and boron had a significant dependency on the fill volume for both filling media. In contrast, the concentrations of aluminum were low with no clear correlation with fill volume for the vials filled with purified water and were below the limit of quantification for the vials filled with 15 wt % KCl solution (see Tables IV and V). Our findings of low concentrations of aluminum in water-filled vials after storage agree with previous observations for borosilicate and aluminosilicate glass containers (3, 7, 9). Because of their greater relevance, the following detailed discussion will focus on the behavior of the dominant leachables: sodium, boron, and silicon.
Standard Quality Vials Filled with Purified Water
The amount of leached sodium in purified water increased from 1.6 µg/mL to 8.4 µg/mL when the fill volume decreased from 2 mL to 0.5 mL in standard quality vials. This means that for the lowest fill volume (0.5 mL) a more than five times higher concentration of sodium was found (factor increase of 5.3). We have added the limit of the hydrolytic resistance test for sodium (2.4 µg/mL) in Table IV for reference. The information is only indicative as the resistance test conditions were 1 h autoclaving at 121°C with the vial filled to 90% of the brimful volume (8), which differ from the conditions of the leaching study.
A similar behavior was observed for boron as depicted in Figure 1 and summarized in Table IV. The boron concentrations found in purified water after 24 weeks of storage in standard quality vials increased from 0.7 µg/mL to 4.1 µg/mL when the fill volume was reduced from 2.0 mL to 0.5 mL or were almost six times higher (factor increase of 5.9).
Concentrations (μg/mL) of leached Na and Si for standard quality, low-fill quality, and aluminosilicate vials and concentrations of B for standard quality and low-fill quality vials filled with different fill volumes of purified water after 24 weeks of storage at 40°C.
For sodium and boron, the increase in concentrations seen when the fill volume was reduced by far exceeded the increase expected because of geometric considerations revealing an increased leachable propensity of the heel zone. “Heel zone” is understood to be the small wall area close to the bottom of the vial where a high-temperature gradient was present during the conversion process of the vial (in this study about 2 mm above the bottom). The term “heel zone” is a synonym for the term “wall near the bottom area.” This is in accordance with the diffusion of sodium and boron into the heel zone during the converting process described elsewhere (4, 6), which leads to a sodium- and boron-rich near-surface microzone with low durability. It has been shown (4) that this zone can extend into the glass for about 100 nm.
For standard quality vials, 15 µg/mL silicon was found in solution at a 0.5 mL fill volume, which is three times more than the 5 µg/mL found in solution for a 2 mL fill volume. This increase in concentration is not as pronounced as the increases found for boron and sodium but is significantly greater than the increase calculated based on the change in the wetted area to fill volume ratio. For silicon, the increased leachable concentration is probably attributable to a combination of a pH shift in the purified water and a modified microstructure of the glass surface in the heel zone.
In a first step, the sodium ions (Na+) in the glass exchange with the hydronium ions (H3O+) in the water. This leaching process is obviously more pronounced in the heel zone with its reduced durability that presumably weakens the glass network in that area (defects of bonds and internal interfaces). In addition, a pronounced dissolution of the sodium borate-based glass phase in the heel zone will occur. As a result, the glass interior develops a microstructure with an increased effective surface and therefore less resistant Si-bonds as seen in several studies (6, 10–11). Because of the increased sodium concentration in the solution, the pH value can increase which fosters a nucleophilic attack by the hydroxide ion (OH−) that disrupts silicon–oxygen bonds in the glass matrix (11, 12⇓–14).
A pronounced sodium leaching for lower fill volumes corresponds to a pronounced ion-exchange with hydronium and a pH increase as reported by Kucko et al. (7) This favors an increased nucleophilic attack (dissolution) of the glass matrix leading to higher silicon concentrations.
Low-Fill Quality Vials Filled with Purified Water
The concentrations of sodium, boron, and silicon found in solution in the low-fill quality vials with 2 mL fill volume are slightly lower than the values observed in solution for the standard quality vials with 2 mL fill volume. However, a fundamental change is seen when the fill volume is reduced. For example, boron and sodium concentrations in solution for the standard quality vials were increased >5-fold when the fill volume was decreased from 2 mL to 0.5 mL. In contrast, for low-fill quality vials, only moderate increases in boron and sodium concentrations in solution were found. In fact, the concentrations doubled which is not much greater than the 1.63-fold increase expected from the increased wetted area to fill volume ratio, indicating that the leachable propensity, especially in the heel zone, is significantly improved.
Specifically (see Figure 1 and Table IV), the amount of leached sodium increased by a factor of 1.9 (from 1.4 µg/mL to 2.7 µg/mL) when the water fill volume decreased from 2 mL to 0.5 mL for low-fill quality vials. A similar behavior was observed for boron with an increase in concentration by a factor of 2.2 (from 0.5 µg/mL to 1.1 µg/mL).
The amount of silicon in solution increased only by a factor of 1.23 (from 4.3 µg/mL to 5.3 µg/mL) when the water fill volume was decreased from 2 mL to 0.5 mL, a fact that is surprising at first glance because it is less than expected based on geometric considerations. As previously mentioned, dissolution processes drive the release of silicon. Therefore, a straightforward conclusion would be that the average silicon dissolution of the bottom and the heel zone — the area wetted by the 0.5 mL fill volume — is reduced compared to the unaltered cylindrical wall. Obviously, there is no corrosion-induced microstructure that leads to weaker Si bonds in the heel zone as described for the standard quality vials. In addition, the near-surface region of the bottom features a fire-polished morphology and a silica-enriched composition because of the evaporation of sodium and boron species during hot forming (4). As a result, the bottom behaves more like fused silica with significantly lower dissolution rates in water compared to borosilicate glasses (12), leading to a small or negligible contribution to the observed silicon concentrations. Of course, this also applies for standard quality vials, but the high leachable propensity in the heel zone dominates in those vials.
Aluminosilicate Vials Filled with Purified Water
Use of aluminosilicate glass for pharmaceutical packaging has been recently discussed (9). Until now, not much data about the typical leaching behavior of aluminosilicate glass has been available in the literature. The vials made of aluminosilicate glass filled with 2 mL solution had about twice the amount of leached sodium (2.8 µg/mL) as the vials made of borosilicate glass (1.4 µg/mL for low-fill quality vials and 1.6 µg/mL for standard quality vials), whereas the silicon concentration was more or less the same for both glass types. It seems obvious that this behavior is mainly caused by the glass composition. Alkali aluminosilicate glasses typically contain >10 wt % of sodium oxide as a requirement for ion-exchange processing in order to improve their surface strength but with the drawback of high sodium leaching (unstrengthened glass) or high potassium leaching (strengthened glass). In this context, it is clear why parenteral packaging containers based on aluminosilicate glass are surface treated (9) to achieve the Type I specification.
Lowering the fill volume of the aluminosilicate vial from 2 mL to 0.5mL resulted in an increase by a factor of 2.5 in the sodium concentration (from 2.8 µg/mL to 6.9 µg/mL) and by a factor of 1.6 in the silicon concentration (from 4.7 µg/mL to 7.3 µg/mL). The increase of the sodium concentration exceeded the 1.63-fold increase expected from the geometric factor indicating an increased leaching propensity in the areas affected by the converting process. In contrast, the behavior of the silicon concentration can be explained by a homogenous dissolution rate for the entire inner surface.
For parenteral applications, it is convenient to express the leachables amounts in weight per volume (e.g., µg/mL) and our discussion of the results follows that path. However, it may be easier to understand the behavior seen for the different fill volumes when the influence of the material-related factor is viewed separately. Such an approach leads to the amount of leachables in weight per area of wetted surface (e.g., µg/cm2) as depicted in Figure 2 for sodium and silicon concentrations after storage of vials filled with purified water for 24 weeks at 40°C.
Concentrations (μg/cm2) of leached Na and Si for standard quality, low-fill quality, and aluminosilicate vials filled with different fill volumes of purified water after 24 weeks of storage at 40°C.
The two diagrams in Figure 2 illustrate that the sodium leachable propensity of the low-fill quality vials is low and more or less independent of the fill volume, whereas the standard quality vials and the aluminosilicate vials had a different behavior, resulting in three-to-four times higher values for a 0.5 mL fill volume. In Figure 2, the silicon leachables per area are nearly identical for all vial types with a 2 mL fill volume. At lower fill volumes, the leachables value does not significantly change for the aluminosilicate vials, whereas an increase of 75% for the standard quality vials and a decrease of ∼25% for the low-fill quality vials was seen at 0.5 mL.
Low-Fill and Standard Quality Vials Filled with 15 wt % KCl Solution
The concentrations of sodium, silicon, and boron found in 15 wt % KCl solution with all three fill volumes and after storage in both vial types are listed in Table V and depicted in Figure 3. Compared to the results derived for purified water, the concentrations observed for low-fill quality vials are obviously lower for silicon and boron and also for sodium if the sodium blank value (Table III, 1.3 µg/mL) is considered as an offset. An opposite behavior is seen for the standard quality vials, as all leachables levels are higher than those observed for purified water. As a result, tremendous differences are found with decreasing fill volume, leading to approximately nine times higher concentrations for boron, five times higher concentration for sodium and approximately 17 times higher silicon concentrations for the standard quality vials compared to the low-fill quality vials.
Concentrations (μg/mL) of leached Na, Si, and B for standard quality and low-fill quality vials filled with different fill volumes of aqueous 15 wt % KCl solution after 24 weeks of storage at 40°C.
Accelerated lifetime studies at 60 °C with 15 wt % KCl solution (4) gave an indication of strong glass attack in the heel zone of standard quality vials that was associated with a pronounced change in the microstructure of the interior surface. The mechanisms leading to the microstructure for the standard quality vials have already been discussed and it seems reasonable that an ion-exchange between sodium and potassium will further weaken the Si–O–Si network and increase the leachable propensity and the dissolution of silicon.
In contrast, a remarkably low amount of dissolved silicon was found for the low-fill quality vials filled with 15 wt % KCl compared with purified water (approximately four times lower). Previous studies found a decreasing influence of dilute KCl solution on the pH value change during storage at 60°C, leading to the conclusion that KCl inhibits the exchange of sodium with hydronium ions (15). Applying this conclusion to our findings, KCl will prevent a pH shift toward a more alkaline range, thereby lowering the hydrolytic dissolution of silicon, which preferentially takes place in alkaline solutions.
Conclusion
A significantly different impact of the fill volume on the leachables profile was seen for the three vial types under investigation. For the standard quality borosilicate vials, the wall near the bottom area is more susceptible to dissolution leading to a strong increase of leached “glass” elements with low fill volumes of water or salt solution. Obviously, an alteration of the chemical durability occurs within the heel zone of the vials during standard converting procedures. In contrast, the study indicates that using automated 100% control over the relevant process parameters and a statistical process control for verification (Quicktest) can potentially reduce the vulnerably of tubular borosilicate vials. The resulting low-fill quality vials feature high resistance to dissolution and in consequence low concentrations of leached “glass” elements when low fill volumes are used. In particular, the reduction in sodium release could be very beneficial for unbuffered product solutions as it prevents a shift in pH value and conductivity.
In general, reduced concentrations of leachables will reduce the interaction with the active pharmaceutical ingredient or the excipients, which might be required for sensitive biologics or for ophthalmic applications where low fill volumes are typically used.
Conflict of Interest Declaration
The authors are employees of the SCHOTT Group which manufactures and sells glass vials and which provided the vials used in this study.
- © PDA, Inc. 2019
