Carol Rea Flynn

	Peer-Reviewed
	Submitted: March 31, 2015; Accepted: April 20, 2015.

Glass, even exceptionally inert Type I borosilicate glass, is susceptible to attack by aggressive chemicals, causing it to spall. This phenomenon in primary packaging for parenteral products, which is referred to as delamination, has increased since the introduction of innovative biopharmaceuticals. A long-term study was conducted to determine the causes of delamination. Of the factors in the manufacturing process that can affect the propensity for vial delamination, this study identified the forming temperature of the vial bottom as the greatest correlation to the phenomenon. This knowledge enabled development of an improved manufacturing process for glass vials and a new type of vial.

	Glass is recognized as the gold standard for primary packaging of parenteral drugs. Glass is durable, inert, clean, and transparent-properties that are important in the manufacture of containers for sophisticated applications. The composition of the glass used has been continuously improved over time. With its high hydrolytic resistance, borosilicate glass (Type I glass) has long replaced the simple soda-lime glass for the packaging of parenteral products.

Nonetheless, even a material as resistant as Type I glass can be tested to the limit by aggressive chemicals. When packaging liquids have a particularly high or low pH value, for example, an interaction with the glass can take place in which the entire surface of the glass corrodes due to the migration of alkali ions from the glass into the solution. Another known phenomenon is when layers of glass flake off in specific zones of the packaging product and contaminate the solution. This process, which is called delamination, is observed during the long-term storage of high-pH solutions and solutions containing chelating agents or buffer systems as well as during the storage of biologics such as monoclonal antibodies. 

 shows an extreme case of delamination created in the laboratory, not under realistic pharmaceutical storage conditions. A series of product recalls in the US has affected all the major manufacturers of primary pharmaceutical packaging made of glass. Delamination has become a serious problem from both a medical and an economic viewpoint. The detached glass flakes could potentially lead to health problems in conjunction with intravenous or intramuscular injection, although no cases of this kind are known yet. In addition, each product recall generates significant losses for the pharmaceutical industry, especially in the high-cost biologics sector.


	In view of the increasing number of product recalls, in early 2011, FDA instructed the pharmaceutical industry, in collaboration with primary packaging manufacturers, to come up with solutions for the delamination problem. Joint meetings involving representatives of the industry and the regulatory bodies were held to assess the current state of knowledge and discuss the future course of action. One year later, on this basis, the US Pharmacopeial Convention (USP) presented a new draft of General Chapter <1660> proposing test procedures going beyond those previously laid down (1). The use of these more aggressive test conditions is intended to facilitate more reliable predictions of the tendency of primary packaging to delaminate.

	In spite of these efforts, a host of questions remained unanswered on the subject of delamination. Little was known about the causes of the phenomenon at the time, and there was no established industry standard for methods of systematic investigation.  A comprehensive long-term study on the subject of delamination was conducted in association with Alfred University (New York, USA). The first task was to develop a reliable method for the qualitative and quantitative analysis of delamination. The study then forced delamination under laboratory conditions to identify the causes of this phenomenon.

 Five different solutions (highly purified water; sodium chloride [NaCl] solution with citrate buffer pH 7; and solutions with pH values of 5.5, 8, and 9.5) were used in the study. The solutions were tested for the presence of visible and subvisible particles, extractables (i.e., chemical compounds and inorganic elements), and pH changes. The identification of particle formation in the solution using dynamic light scattering (DLS) proved to be inconclusive. So, in addition to pH determination by µ-electrode, the study focused in particular on the filtration of the solution with subsequent energy-dispersive X-ray spectroscopy (EDX) of the filter. Extractables were analyzed using an inductively coupled plasma (ICP) and micronebulizer. A hydrolytic resistance test in accordance with 

 <660> was also conducted (2, 3). This test determines the resistance of the interior surfaces to migration of ions out of a filled vial.

	Data were also collected on the vials. Surface morphology data collected by scanning electron microscope (SEM) was the main method used over and above visual inspection. Stress in the section of the vial particularly at risk to delamination was also visualized using a polariscope. A methylene blue surface-staining test was conducted, but was only of secondary significance.

 Once a reliable analytical procedure had been developed as described in the previous section, a design of experiment (DOE) study was created to identify the root cause(s) of delamination. To ensure that all the potential influence variables were taken into account, a broad approach was chosen encompassing the relevant factors during vial production and also during the subsequent pharmaceutical processes, which are illustrated in 

. First, in the area of material formulation, the two borosilicate glass types (Type I class A Gx33 and class B Gx51-V) were tested. For control purposes, a batch of vials made of molded glass was included in the study. In each case, 3-mL and 10-mL vials were investigated. During the production process, the effects of different burner settings and the associated forming temperatures were determined. As both temperatures and temperature gradients are high during formation, these parameters were the prime focus of the study. The effects of three temperature steps during the formation of the vial bottom were compared, and the potential influence of continuous and indexing forming processes was investigated. A further focus of the study was the effect of treating the vials with ammonium sulfate to remove alkali ions from the surface (pre-leaching) and, in this way, render them more hydrolytically resistant. On the processing side, the effects of washing, sterilization, and depyrogenation were also investigated.

	Forty-eight batches with an approximate total of 100,000 vials were produced and delivered to Alfred University for the study where they were filled with test solutions and stored at 5, 25, and 40 °C. The degree of delamination was investigated at the beginning of the test and again after three months, six months, and one year. Further examinations will take place after three years.

	On conclusion of the evaluations made at the end of the one-year storage period, an additional series of tests using USP <1660> was conducted with correspondingly higher solution concentrations and temperatures as follows (1):

		0.9% potassium chloride (KCl), pH 8.0, 2 h at 121 °C

		20 mM glycine, pH 10.0, 24 h at 50 °C.

	For these additional vials, an evaluation was made after 2 or 24 hours (as required by 

 After one year, detailed results on the triggers and course of delamination were seen. The destruction of the glass surface to the point where particles become detached is clearly a time-dependent phenomenon. Delamination was more evident the longer the glass came into contact with aggressive solutions. Corroded surfaces and detached glass flakes occurred with increasing frequency during later tests (four to six months and longer).

	The changes observed during the study under realistic storage conditions were confirmed during the test series in accordance with 

<1660>, which involves more stringent test conditions. Despite high solution concentrations and extreme temperatures, no corrosion or detached glass flakes were observed during the test phase (2–24 hours), with glass flakes only becoming detached after several months, as shown in an SEM in 

 There were also differences in the specific effects of different solutions. In conjunction with Gx51-V glass, KCl solution had the most pronounced delamination effect, followed by solutions with citrate buffer and glycine. With Gx33 glass, however, the solution with citrate buffer had the most pronounced delamination effect, followed by solutions with glycine and KCl.

 Treating the vials with pre-leaching makes the surface of the glass susceptible to delamination. The most severe forms of delamination were found in conjunction with treated glass. Taking all sample types together, filled with aggressive solutions, treated vials showed statistically significant higher occurrence of delamination than untreated vials. Treated glass should therefore be avoided when packaging delamination-inducing medicinal products, no matter how successful it has proven to be in other areas of application.

 An interesting finding in this study is that in the manufacturing process, the factor that has the greatest correlation to the propensity for delamination is the bottom forming temperature. It was observed that only the heel area of the sample vials delaminated, and this area is subject to particularly high thermal loads.

, vial delamination typically occurs in a clearly defined region of interest (ROI) between the heel stress ring and the base of the vial due to the base forming process. In this process, the vial is turned upside down and the upper part is heated to such a high temperature that sodium, potassium, and boron evaporate from the glass matrix. The vapors released condense on cooler parts of the vial. At the same time, the temperature conditions directly underneath the base permit the formation of a secondary thin layer of glass with an average thickness of 0.5 µm. This process is abetted by the supply of sodium and potassium in the vapor, which lowers the melting point. In contrast, the body of the vial is so cool that no new layer of glass can form. The surface layer of glass at the stress ring is less inert than the tubular glass and has a different thermal expansion coefficient, which makes it more susceptible to attack and degradation by stored solutions. The central role of the forming temperature in delamination also explains the initially surprisingly result that the 10-mL vials were more prone to delamination than the 3-mL vials. Larger vials generally have greater wall thicknesses, necessitating a correspondingly greater supply of heat during formation, which resulted in a greater tendency to delaminate.


	This finding that vial delamination is primarily triggered by the temperature conditions during vial formation, in addition to the interaction of the glass with the pharmaceutical product, makes it possible to systematically improve the production process. To produce vials intended for the storage of innovative or particularly aggressive pharmaceutical drugs, formation below a defined temperature limit takes priority over other factors, such as optical perfection of the vial bottom. By implementing quality-by-design principles in the production operations, the input parameters-in this case, the temperature of the bottom forming burners-are continuously adjusted according to the output on the basis of statistical process control (SPC).

	If the evaporation and recondensation of sodium, potassium, and boron from the glass matrix is the most significant cause of delamination, any optimization of the production process must start at precisely this point. The most effective way to achieve this is to optimize the forming temperature, especially when smoothing the base of the vial. During the study, vials were produced across the entire forming temperature spectrum. The cumulative results across all other DOE variables show a clear correlation between the burner temperature and the number of delaminated vials, as shown in 

. Here, different process-related temperature ranges were taken into account for continuous and indexing machines. However, the correlation applies equally to both production methods. It is, therefore, possible to determine temperature limits for each machine type below which delamination in the ROI is greatly reduced. This being said, there are still aggressive solutions that are not recommended to be used in Type 1 glass containers.

The THOR process (Thermal Hydrolytic Optimization and Reduction) was developed by Gerresheimer to limit the process temperatures and use an improved bottom heat profile. A decisive factor for the success of the optimization is that THOR is designed as a feedback SPC-controlled cycle.

	In the process cycle, the forming temperature (entire bottom matrix and not a point sensor) of every single vial is measured inline by camera systems using proprietary software for the cameras and subsequent data processing. The results are forwarded to the Infinity SPC system on the production lines, which rejects all vials where the forming temperature was exceeded. In addition, the method of displaying process statistics at the hot end was modified so that an on-screen warning signal is generated on occurrence of excessive forming temperatures. The forming parameters can be adjusted in this case to keep the process within the preset limits and reduce the reject rates. Another positive side effect of THOR cycle production is the improved hydrolytic resistance of the vials. The long-term study shows that vials produced within the defined formation temperature limits fail the hydrolytic resistance test far less often.


	A proprietary test procedure (Gerresheimer’s FLASH test) takes into account the specific occurrence of delamination in the stress ring near the bottom of the vial. In this 

 3.2.1.A-compliant test, a vial is filled with test liquid up to the nominal volume and, after autoclaving, the solution is titrated to determine the hydrolytic resistance of the vial. If, however, the tendency for delamination of the vial is the focus of the test, it makes sense to fill the vial only to a level just above the ROI in order to obtain precise, delamination-specific test results. 

 shows a high correlation between the test results with a reduced filling volume and the rate of delamination of vials in the framework of the study for two types of glass (Gx33 and Gx51). This method can be used as a quality-control test.


	As there is no reliable way to replace long-term testing, the study will be continued for another year to deliver further results on the interaction of stored pharmaceuticals and vials produced under different conditions.

 General Chapter <1660>, “Evaluation of the Inner Surface Durability of Glass Containers” (US Pharmacopeial Convention, Rockville, MD, 2012).
	2. 

 3.2.1A, “Glass containers for pharmaceutical use” (EDQM, Strasbourg, France, 2013).
	3. 

 General Chapter <660>, “Container–Glass” (US Pharmacopeial Convention, Rockville, MD, 2013).


	Carol Rea Flynn is director of technical services, Dan McNerney is director of engineering, and Palak Shah is engineering specifications manager, all at Gerresheimer, Vineland, NJ. Address all correspondence to Marion Stolzenwald, M.Stolzenwald@gerresheimer.com.

:
	When referring to this article, please cite it as C.R. Flynn, D. McNerney, and P. Shah, "Identifying Causes of Delamination," 

Articles

Delamination of glass packaging is a source of particulates in parenteral drugs, but identifying the root cause allowed the design of an improved manufacturing process for glass vials.