`
`
`
`Downloaded from on April 13, 2020Downloaded from on April 13, 2020
`
`Accelerated Extractable Studies of Borosilicate Glass
`Containers
`Steven J. Borchert, Michele M. Ryan, Richard L. Davison, et al.
`
`
`
` PDA J Pharm Sci and Tech
`
`1989
`
`43,
`
` 67-79
`
`Opiant Exhibit 2323
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685, IPR2019-00688, IPR2019-00694
`Page 1
`
` 
` 
`  
`

`

`• RESEARCH
`
`ARTICLE
`
`Downloaded from
`
`on April 13, 2020
`
`Accelerated Extractable Studies of Borosilicate Glass Containers
`
`STEVEN J. BORCHERT*A, MICHELE M. RYAN*, RICHARD L. DAVISON*, and WILLIAM SPEED*
`
`* The Quality Control Division, The Upjohn Company, Kalamazoo. Michigan. f The Quality Control Division. The Upjohn Company,
`Crawley, United Kingdom
`
`ABSTRACT: This article describes the use of an accelerated extractable procedure for borosilicate glass
`containers. The procedure, which is very similar to a protocol developed by a PDA Task Force, includes the
`monitoring of Si. Na. K. Al. Ba. Ca. Mg, Fe. and Zn in the extracts as well as measurements ofpH change and
`total extractables. Unlike the PDA protocol, which uses HiO as the sole extraction solution, the procedure
`outlined in this report used a variety of unbuffered (pH ^ 4, H:0. pH ¡Ü 6.5. pH ^ 8.0. pH m 9.5, andpH =*
`¡0.4) and buffered (pH — 8 and pH = 10) aqueous extraction media. Studies were completed for several
`borosilicate glasses, including a mixture of tubing vials, molded vials, and ampoules from US and European
`suppliers. Results of these studies are presented in this article and are discussed in terms of the interactions
`between borosilicate glasses and aqueous solutions.
`
`Introduction
`
`Glass containers used in the pharmaceutical industry
`are durable, but cannot be considered inert. When it is
`placed in contact with a solution, a variety of interactions
`can occur, including ion exchange or selective leaching,
`glass dissolution, pitting, solution concentration, precipi(cid:173)
`tation, stable film formation, surface layer exfoliation,
`weathering, stress corrosion, and erosion corrosion ( I ).
`Glass/media interactions have been studied over sixty
`years, and during this time period several hundred arti(cid:173)
`cles, numerous reviews (1-9), and two extensive bibliog(cid:173)
`raphies (10-11) have been published.
`Glass corrosion has been a very important subject in a
`variety of fields outside the pharmaceutical industry ( 1-
`34). For example, the weathering of glass has been exten(cid:173)
`sively studied to determine if glass processing costs can be
`reduced without sacrificing durability and to develop opti(cid:173)
`mum methods for preserving historic windows and glass
`treasures. The wide use of glass electrodes for chemical
`analyses and glass columns for chromatographic separa(cid:173)
`tions has promoted a detailed characterization of glass/
`media interactions. Finally, the leaching of glass and oth(cid:173)
`er glass corrosion mechanisms have been studied in rela(cid:173)
`tion to the storage of radioactive wastes as glasses and the
`use of glass fiber optics for communication purposes.
`Numerous kinds of glasses, types of media, and expo(cid:173)
`sure conditions have been used in studies from these other
`disciplines (1-34). Borosilicate glasses (USP Type I),
`soda-lime glasses (USP Types II and III), fused quartz.
`Vycor glass, simulated nuclear waste glasses, and binary
`
`Received Mav 13, 1988. Accepted for publication November 15.
`1988.
`4 Author to whom inquiries should be directed.
`
`or ternary glass mixtures of oxides have been examined.
`The medium used for most studies was H20. although
`aqueous solutions containing various salts or chelating
`agents, acidic media, and alkaline solutions have also been
`used. In terms of the environmental conditions reported in
`these studies, glasses were exposed to the media for long
`time intervals (2-3 years) at room temperature (25°C) or
`for significantly shorter time intervals ( 1 hour to 30 days)
`at elevated temperatures (70-200°C).
`The results of weight loss measurements, solution ana(cid:173)
`lyses, and various surface analytical measurements from
`studies in nonpharmaceutical fields (1-34) have led to a
`good understanding of the qualitative and quantitative
`characteristics of glass/media interactions. In addition,
`some of these studies have also provided the scientific
`basis for the different water attack, glass extraction, and
`powdered glass tests used by the parenteral industry (4).
`Glass and its interactions with solutions have also been
`important topics in the pharmaceutical industry. Al(cid:173)
`though the pharmaceutical bibliography represents only a
`small subset of the overall glass bibliography from all
`disciplines, several articles (35-40) have reviewed the ma(cid:173)
`terial in relevant pharmaceutical papers. In general, both
`review and individual articles have listed the composition
`of pharmaceutical glasses (36-38, 40-42), have discussed
`the types of interactions between glass containers and
`pharmaceutical solutions (35-41, 43-48), or have de(cid:173)
`scribed the various protocols for glass testing (39-40, 44-
`46).
`Concerning the composition of pharmaceutical glasses,
`a PDA report on Glass Containers (37), Technical Meth(cid:173)
`ods Bulletin No. 3, described the classification of paren(cid:173)
`teral glasses according to the USP classification system
`and the glass manufacturers' designations. Representa(cid:173)
`tive compositions of common pharmaceutical glasses were
`
`Vol. 43. No. 2 / March-April 1989
`
`67
`
`Opiant Exhibit 2323
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685, IPR2019-00688, IPR2019-00694
`Page 2
`
`

`

`Downloaded from
`
`on April 13, 2020
`
`also provided in this report. In addition, this Technical
`Methods Bulletin contained a discussion of factors that
`can be used for the selection of a glass container and a
`review of identification test methods for glass containers
`used by the parenteral industry.
`In the case of glass/product interactions many authors
`have taken the knowledge about glass/media interactions
`acquired from other fields and have directly applied the
`information to glass compositions and media used in the
`pharmaceutical industry. One aspect of glass/product in(cid:173)
`teractions, glass extractables, has been the focus of several
`papers. The nature of the extractables, the mechanisms by
`which they arise, and the factors affecting them have been
`the topics for discussion in these articles (35, 37, 39. 41,
`43, 45, 47). However, drug absorption and other glass/
`product interactions have also been thoroughly discussed,
`especially in the review by Wang and Chien (38).
`As mentioned above, several extraction procedures
`have been developed for pharmaceutical containers (39-
`40, 44-48). Similar to glass extractable experiments used
`in other disciplines, the medium used in most of the proto(cid:173)
`cols is H20. In addition, most procedures utilize acceler(cid:173)
`ated extraction conditions where the glass containers are
`exposed to the extraction solution at an autoclave tem(cid:173)
`perature (121°C) for relatively short time intervals (1-20
`hours). However, extraction conditions using lower tem(cid:173)
`peratures (25-70°C) for significantly longer time inter(cid:173)
`vals (3-4 months) have occasionally been used (47). Fi(cid:173)
`nally, the types of extractable measurements have in(cid:173)
`volved total extractable alkali, pH shifts, and specific
`elemental analyses.
`Among the various extraction procedures used for
`pharmaceutical containers, the PDA Extractable Proto(cid:173)
`col (39) is probably the most general. It is very similar to
`the USP Water Attack test (49) with one major excep(cid:173)
`tion. It uses the same extraction medium, H20. and it
`subjects the containers to the same exposure conditions.
`121°C for one hour, as that of the compendial test. How(cid:173)
`ever, the protocol includes the monitoring of several ele(cid:173)
`ments such as Na, Al, K, Ca, Ba. Fé, Mg, Zn, and Si in the
`extracts. In addition, the procedure includes measure(cid:173)
`ments of the pH change and the extractable weight and
`suggests methods of identifying the materials leached
`from glasses. In contrast, the USP Water Attack test
`contains only one measurement, a pH titration which is
`indirectly related to extractable alkali.
`It is particularly noteworthy that water is the only
`extraction medium specified in the PDA Extractable Pro(cid:173)
`tocol. However, the PDA Task Force on Glass Extractives
`noted that the procedure, in principle, could be used with
`other extraction solutions although some of the tests such
`as pH change, extractable weight, and the monitoring of
`certain elements may not be applicable for some extrac(cid:173)
`tion media (39). In our glass extractable experiments we
`have used the PDA protocol, but we have used a variety of
`unbuffered (pH =* 4. H20, pH =* 6.5, pH =* 8.0, pH =*
`9.5, and pH ^ 10.4) and buffered (pH = 8 and pH = 10)
`aqueous extraction media. This report contains the results
`of our studies with many different US and European
`borosilicate parenteral glass containers.
`
`Experimental
`The extraction studies were done in two different lab(cid:173)
`oratories, one in Crawley. UK and the other in Kalama(cid:173)
`zoo. MI. Similar experimental procedures were used in
`both laboratories with the exception of the methods used
`for some of the elemental analyses.
`
`Preparation of Extraction Media
`In addition to High-Purity H20 (e.g., Milli-Q® Grade
`Water), which is the extraction medium specified in the
`PDA protocol (39), other unbuffered (pH m 4. pH a¡ 6.5.
`pH « 8.0, pH en 9.5. and pH s* 10.4) and buffered (pH
`= 8 and pH = 10) aqueous extraction media were also
`used. With the exception of water, the unbuffered systems
`were prepared by adjusting the pH of a High-Purity H20
`solution using a dilute solution of hydrochloric acid ( Ana-
`chemia; AristaR) or lithium hydroxide (Alpha Products:
`AristaR). The pH = 8 buffer was prepared by adjusting
`the pH of a 0.01M solution of Tromethamine (2-amino-2-
`hydroxymethyl-l,3-propanediol); whereas, the pH = 10
`buffer was a 0.05vW sodium carbonate/0.05M sodium
`bicarbonate solution.
`The water and the reagents used to adjust the pH were
`chosen so that the elemental concentration in each unbuf(cid:173)
`fered extraction medium was < the appropriate detection
`limit for each of the nine elements analyzed. Similar pre(cid:173)
`cautions were taken for the preparation of the buffered
`media, and the concentration of Si was < detection limit.
`The detection limits for each element are noted in a later
`section of this report.
`
`Preparation of Samples
`
`All containers were treated as specified in the PDA
`protocol (39). In particular, all vials and ampoules were
`rinsed twice with High-Purity H20.
`
`Preparation of Standard Solutions
`
`An aqueous standard stock solution of certified purity
`was acquired for each element from a commercial source
`(Spex Industries Inc., National Bureau of Standards, or
`Scientific Products). For example, the standard stock so(cid:173)
`lution for Na was a 1000-ppm Na solution that was ob(cid:173)
`tained from Spex Industries Inc. In general, the elemental
`concentration of each stock standard solution was approx(cid:173)
`imately 1000 ppm for the element of interest.
`As specified in the PDA protocol (39), working stan(cid:173)
`dard solutions were prepared fresh daily from the stock
`solutions and were stored in acid-rinsed (1 + 1 HNO?)
`polyethylene bottles. With the exception of Mg, the work(cid:173)
`ing standard solutions were very similar to those noted in
`the PDA protocol (39). In the case of Mg, the working
`standard solutions were prepared using 0.6% La.
`
`Extraction Procedure
`
`The extraction procedure was very similar to the one
`specified in the PDA protocol (39). First, the pH of each
`extraction medium was measured just prior to the experi(cid:173)
`ment. Before measuring the pH, 10 mg of reagent grade
`KC1 was added and dissolved in a 10 mL aliquot of the
`extraction medium.
`
`68
`
`Journal of Parenteral Science & Technology
`
`Opiant Exhibit 2323
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685, IPR2019-00688, IPR2019-00694
`Page 3
`
`

`

`Downloaded from
`
`on April 13, 2020
`
`Next, the rinsed sample containers were filled to their
`nominal capacity with the extraction medium. A mini(cid:173)
`mum of twenty vials or amoules was used for each experi(cid:173)
`ment. More than twenty of the smaller volume containers
`were used in order to obtain a volume of 200 mL. which
`was necessary to complete the various analyses described
`below.
`After all the containers were filled, each was covered
`with a polypropylene (PP) beaker. The PP beakers, which
`were used as covers, had been previously rinsed twice with
`High-Purity H2O and had been exposed to H20 under
`autoclave conditions (121°C for 1 hour).
`Finally, the filled and covered containers were auto-
`claved for one hour at 121 ± 1°C. After the containers
`were cooled, the contents of all glass containers were
`emptied into an acid-rinsed (1 + 1 HNO3) polyethylene
`bottle.
`
`Analyses of Extracts
`
`The combined extracts were subjected to several mea(cid:173)
`surements. A summary of the analytical procedures used
`in the US studies, which were very similar to those report(cid:173)
`ed in the PDA protocol (39), is given below.
`a. pH: The pH of each extract was measured immedi(cid:173)
`ately after the extracts were pooled together. Before mea(cid:173)
`suring the pH, 10 mg of reagent grade KC1 was added and
`dissolved in a 10 mL aliquot of the extract.
`b. Elemental Analyses: For the analyses of Na. K. Fe.
`and Zn, an aliquot of the sample extract was analyzed
`using the appropriate atomic absorption (AA) procedure.
`The concentrations were determined using an air/acety(cid:173)
`lene flame and the following AA bands: Na. 589.0 and
`589.6 nm peaks; K. 766.5 m band; Zn, 213.9 nm peak: and
`Fe, 248.3 nm band. Dilution of some samples was neces(cid:173)
`sary in order to perform some of the measurements in a
`suitable concentration range.
`For the AA determinations of Ba and Al. each extract
`was analyzed in the presence of 1500 ppm NaCl. A blank
`was prepared in the same manner. The concentrations
`were determined using a nitrous oxide/acetylene flame
`and the appropriate AA bands (Al, 309.3 nm; Ba. 553.6
`nm). All the Al and Ba measurements were obtained using
`scale expansion (10X).
`Ca and Mg were analyzed in the presence of 0.6% La. A
`blank was prepared in the same manner. The concentra(cid:173)
`tions were determined using an air/acetylene flame and
`the appropriate AA bands (Ca. 422.7 nm; Mg, 285.2 nm).
`The modifications to the Heteropoly Blue Method, not(cid:173)
`ed in the PDA protocol (39). were used for the analysis of
`Si (as SiOi). The band at =±650 nm was used for all the
`measurements. Also, an extra step, the addition of 0.1 mL
`of 1 + 1 H2SOj prior to the addition of the 1 + 1 HC1 to
`the solutions, was used in the Si determinations of the pH
`=± 10 buffered media.
`c. Miscellaneous Measurements: Each extract was mea(cid:173)
`sured for total extractables by transferring a 100 mL
`aliquot of the extract to a tared evaporating dish, evapor(cid:173)
`ating the solution on a steam bath to dryness, drying the
`dish in an oven at 105°C for one hour, cooling it in a
`desiccator, and weighing it. A blank was treated in the
`
`same manner using extraction medium that had not been
`exposed to glass containers. The total extractables. in mg
`per 100 mL of solution, was reported as the difference
`between the amount found in each extract and the blank.
`Some of the sample and blank residues from the measure(cid:173)
`ments for total extractables were also analyzed by X-ray
`fluoresence and infrared spectroscopy.
`The analytical procedures used in the UK studies were
`the same as those described above with some noteworthy
`exceptions. In the UK experiments, analyses for Al, Ba,
`Ca, Mg, Fe. Zn, and Si were performed using graphite
`furnace AA procedures instead of the AA flame methods
`and the Heteropoly Blue Method used in the US studies.
`
`Elemental Analyses: Accuracy, Linearity Range, and
`Detection Limits
`
`Information pertaining to the accuracy, linearity range,
`and detection limit of each elemental assay is essential to
`properly evaluate the results of any extraction study. In
`principle, a number of variables can affect the accuracy of
`any elemental determination. However, since the level of
`extractables found in most glass extractable studies is
`usually small, there was probably only one factor, the
`matrix effect of the extraction medium, that could have
`significantly affected the elemental analyses reported in
`this report.
`To determine if the accuracy of each elemental analysis
`was adequate, standards, which had been prepared at
`equivalent concentrations in water (pH ^ 6.0) and three
`unbuffered aqueous solutions (pH =± 2.0, pH =± 4.0, and
`pH =; 10.0), were analyzed. In these studies, a deviation
`not greater than 10% between the absorbance of standards
`prepared in H2O and those prepared at different pH levels
`was defined as acceptable. The results are shown in Table
`1. For each of the elements that were analyzed, two differ(cid:173)
`ent concentrations of standards were examined. The re(cid:173)
`sults for most elements were considered acceptable. For
`those elements (Mg and Na) for which absorbances were
`not equivalent, some adjustments in the sample prepara(cid:173)
`tions were made to overcome unequal response problems.
`For example, as noted above, all of our Mg determinations
`were made in the presence of 0.6% La. The procedure for
`Mg, which is specified in the PDA protocol (39), does not
`include the addition of La, and this could lead to difficul(cid:173)
`ties if a wide range of pH values were used (see Table I). It
`is also noteworthy that the addition of KC1 to the solutions
`used for Na determinations was not used even though the
`data in Table I indicates such a modification would have
`been desirable. One reason for omission of the KC1 addi(cid:173)
`tion was that none of our glass studies were done for pH =;
`2, where the largest absorbances differences in the Na
`analyses were observed. Another reason is that analytical
`grade KC1 contains a small amount of NaCl, and this
`leads to a significant bias (=¿0.050 absorbance units:
`=¿0.2 ppm) in the data reported for Na. It is true that the
`data could have been corrected for the bias. Indeed, a
`correction was applied to the Na (w/KCl) data shown in
`Table I. However, the correction was comparable to the
`values found for Na in some glass extractable experi(cid:173)
`ments, and, consequently, the uncertainty in the measure-
`
`Vol. 43, No. 2 / March-April 1989
`
`69
`
`Opiant Exhibit 2323
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685, IPR2019-00688, IPR2019-00694
`Page 4
`
`

`

`Downloaded from
`
`on April 13, 2020
`
`TABLE 1. AbsorbanceV;
`.1 Standards at Differ
`ilues of Elcmema
`•cru pH Values
`
`Absorbance
`
`Element
`
`Standard
`(ppm)
`
`H20
`pH =¡ 6.0
`
`pH =¡ 2.0
`
`pH =¡ 4.0
`
`pH =¡ 10.0
`
`Na
`
`Na"
`(w/KCl)
`Mg
`
`Mg*
`( w / L a)
`
`0.048
`0.2
`0.4
`0.093
`0.049
`0.2
`0.097
`0.4
`0.028
`0.1
`0.078
`0.2
`0.053
`0.1
`0.105
`0.2
`0.4
`0.030
`0.116
`1.0
`0.2
`0.011
`0.021
`0.4
`0.0034
`1.0
`0.0078
`2.0
`0.0047
`0.4
`0.0140
`1.0
`0.047
`0.2
`0.4
`0.095
`0.039
`1.0
`0.076
`2.0
`0.095
`1.0
`3.0
`0.280
`" With 1000 ppm KC1. Data corrected for presence of Na in KG.
`* With 0.6% La.
`'With 1000 ppm KC1.
`
`K
`
`Ca*
`
`Al'
`
`Baf
`
`Zn
`
`Fé
`
`S i O:
`(650 nm)
`
`0.086
`0.144
`0.053
`0.099
`0.040
`0.080
`0.058
`0.109
`0.038
`0.121
`0.012
`0.022
`0.0039
`0.0078
`0.0050
`0.0140
`0.048
`0.094
`0.039
`0.074
`0.095
`0.270
`
`0.063
`0.105
`0.050
`0.100
`0.032
`0.077
`0.055
`0.104
`0.037
`0.122
`0.011
`0.022
`0.0038
`0.0077
`0.0050
`0.0140
`0.047
`0.094
`0.038
`0.074
`0.090
`0.265
`
`0.060
`0.099
`0.055
`0.102
`0.018
`0.066
`0.053
`0.103
`0.034
`0.120
`0.011
`0.022
`0.0035
`0.0076
`0.0050
`0.0140
`0.045
`0.094
`0.038
`0.074
`0.100
`0.275
`
`ments of low Na determinations would probably not have
`been significantly better had the KG addition been uti(cid:173)
`lized.
`Si standards were also prepared at equivalent concen(cid:173)
`trations in unbuffered (pH =¡ 8 and pH ^ 10) and buff(cid:173)
`ered (pH ¡=¡ 8 and pH =* 10) aqueous solutions and
`analyzed using the Heteropoly Blue Method. The results
`for Si were acceptable as long as the extra step, the addi(cid:173)
`tion of H2SO4 described earlier, was used for the measure(cid:173)
`ments of the pH =i 10 buffered media.
`The concentration range in which a linear response was
`observed for each element is shown in Table II. The detec(cid:173)
`tion limit for each element depended upon the method
`
`TABLE 11.
`
`Linearity Rani
`gc and Detection
`mental Assays
`
`Limits for Ele-
`
`Linear
`Range
`(ppml
`
`US Studies
`Detection
`Limit
`(ppml
`
`UK Studies
`Detection
`Limit
`(ppm)
`
`Element
`
`Alu
`0-40.0
`=¡0.005
`=¡0.1-0.3
`0-25.0
`=¡0.005
`=¡0.1-0.3
`Ba"
`0-5.0
`=¡0.005
`=¡0.1
`Ca"
`0-0.5
`=¡0.005
`=¡0.01-0.02
`Mg"
`=¡0.06
`0-1.0
`=¡0.06
`Na*
`=¡0.05
`0-2.0
`=¡0.05
`K*
`=¡0.005
`0-5.0
`=¡0.1
`Fe"
`=¡0.007
`0-1.0
`=¡0.02
`Zn"
`0-2.5
`=¡0.01
`=¡0.1
`Si'
`" US Studies: Flame AA: UK Studies: Graphite Furnace AA.
`h US Studies: Flame AA; UK Studies: Flame AA.
`• US Studies: Heteropoly Blue Method; UK Studies: Graphite Fur(cid:173)
`nace AA.
`
`used for the determination and is also shown in Table II.
`In general, the graphite furnace AA methods, which were
`used in the UK studies, had significantly lower detection
`limits than the flame AA procedures.
`
`Discussion
`
`Extractable Protocol
`
`Table III provides a comparison of the key features of
`the extraction procedure used in our studies with those of
`
`T A B LE III.
`
`Accelerated Procedures for Glass Extraciables
`
`USP
`Water
`Attack
`Test
`
`PDA
`Extractable
`Protocol
`
`Extraction
`medium
`
`H ,0
`
`H ,0
`
`Extraction
`conditions
`
`Autoclave
`1 2 1± 1°C
`1 hr
`
`Autoclave
`121 ± 1°C
`1 hr
`
`Extractable
`Procedure,
`This
`Study
`
`H .O
`Unbuffered H:0
`(pH =¡ 4-10)
`Buffered H:0
`(pH = ¡ 8 - 1 0)
`Autoclave
`121 ± 1°C
`1 hr
`
`Extract
`analyses
`
`Total alkali
`
`pH change
`Na, Al, K.
`Ba. Ca, Fe,
`Mg, Zn, Si
`Total extract-
`ables
`ID methods
`
`pH change
`N a . A I, K,
`Ba, Ca, Fe.
`Mg, Zn, Si
`Total extract(cid:173)
`a b as
`ID methods
`
`70
`
`Journal of Parenteral Science & Technology
`
`Opiant Exhibit 2323
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685, IPR2019-00688, IPR2019-00694
`Page 5
`
`

`

`TABLE IV. Collaborative
`Measurement
`
`-One Borosilicate Glass Lot
`Study": PDA Protocol-
`Laboratory 1
`
`Laboratory 2
`
`Laboratory 3
`
`Laboratory 4
`
`Downloaded from
`
`on April 13, 2020
`
`SiOi (ppm)
`Na (ppm)
`K (ppm)
`Al(ppm)
`Ba(ppm)
`Ca (ppm)
`Mg (ppm)
`Fe (ppm)
`Zn (ppm)
`Total extractables
`(mg in lOOmL)
`Initial pH
`Final pH
`pH change
`" See Reference 50.
`
`13.8
`2.98
`0.65
`0.5
`0.1
`ND
`ND
`ND
`ND
`3.4
`
`5.96
`7.75
`1.79
`
`16.7
`2.98
`0.63
`0.5
`0.1
`ND
`ND
`ND
`ND
`3.6
`
`5.96
`7.97
`2.01
`
`12.0
`l."5
`0.88
`1.4
`ND
`0.01
`ND
`ND
`0.01
`
`•Ñ ~
`
`6.9
`8.4
`1.5
`
`12.6
`2.2
`0.92
`1.3
`0.1
`0.01
`ND
`ND
`0.01
`5.3
`
`6.9
`8.3
`1.4
`
`18.2
`3.25
`0.71
`0.8
`ND
`0.03
`0.003
`0.02
`ND
`4.2
`
`5.57
`8.11
`2.54
`
`18.8
`3.23
`0.72
`0.8
`ND
`0.03
`0.003
`ND
`ND
`3.6
`
`5.63
`8.21
`2.58
`
`8.2
`1.47
`0.36
`0.9
`ND
`0.07
`0.02
`ND
`ND
`4.3
`
`5.7
`8.5
`2.8
`
`9.0
`1.56
`0.33
`0.8
`ND
`0.07
`0.02
`ND
`ND
`3.8
`
`5.7
`8.6
`2.9
`
`the USP Water Attack test and the PDA protocol. With
`the exception of the use of several different unbuffered
`and buffered aqueous solutions in addition to H;0 as the
`extraction medium, our procedure was essentially the
`same as the PDA protocol.
`Table IV contains the results of an interlaboratory
`study (50), in which the PDA protocol was tested on one
`lot of a typical borosilicate glass composition. The levels in
`the major extractables (SiO^, Na, and K) varied by a
`factor of 2-3 among the different companies participating
`in the collaborative study. Since the extraction experi(cid:173)
`ments used in our studies were very similar to the PDA
`protocol, one would expect that the interlaboratory and
`intralaboratory precision of our procedure would be com(cid:173)
`parable to that observed for the PDA protocol.
`
`Samples
`
`Table V contains a description of the borosilicate glass
`containers used in our studies. Except for some experi(cid:173)
`ments with one of the glasses only a single lot of containers
`
`was examined for each type of glass. Both ammonium
`sulfate treated and untreated glass containers were used
`and a mixture of tubing vials, molded vials, and ampoules
`were analyzed. Also, it should be emphasized that our
`studies did not encompass all parenteral glass suppliers.
`The seven borosilicate glasses, which were examined in
`the US experiments, were acquired from two different US
`suppliers; whereas, the seven borosilicate glasses, which
`were tested in the UK studies, were obtained from four
`different European vendors.
`Tables VI and VII contain the approximate chemical
`composition of each of the glasses, which was determined
`by X-ray fluorescence (37). A comparison of the glass
`compositions used with parenteral products (37, 42) with
`the information in these tables suggests that the glasses
`used in our studies were representative of the borosilicate
`containers used in the pharmaceutical industry.
`Finally, the glass containers were not processed (i.e.
`washed and sterilized), although all containers were thor(cid:173)
`oughly rinsed with High-Purity HiO before the extraction
`
`TABLE V.
`
`Description of Glass
`
`Containers—US
`
`and UK Studies
`
`Class
`#
`
`Color
`
`Fabrication
`Method
`
`Molded
`Flint
`US-1
`Flint
`Tubing
`US-2
`Flint
`US-3
`Tubing
`US-4
`Flint
`Tubing
`Amber
`Tubing
`L'S-5
`Flint
`Molded
`US-6
`US-7
`Flint
`Tubing
`Flint
`Tubing
`UK-1
`Flint
`Tubing
`UK-2
`Flint
`Tubing
`UK-3
`UK-4
`Amber
`Tubing
`Flint
`Molded
`UK-5
`Molded
`Flint
`UK-6
`Amber
`Molded
`UK-7
`" (NH4);SO.i = ammonium sulfate treatment
`(NH4);S04-D = ammonium sulfate treatment-double
`
`Surface"
`Treatment
`
`Untreated
`Untreated
`(NH4);S04
`(NH4):S04-D
`(NH4):S04
`Untreated
`(NH4);S04
`(NH4);S04-D
`Untreated
`Untreated
`Untreated
`Untreated
`Untreated
`Untreated
`
`Nominal
`Volume
`(mL)
`
`Surface
`Area to
`Volume,
`Ratio
`(cm2/mL)
`
`10
`10
`10
`10
`10
`20
`5
`10
`2
`10
`10
`20
`50
`50
`
`4.0
`2.8
`2.8
`2.8
`2.8
`2.3
`4.0
`3.0
`4.6
`3.0
`3.0
`2.0
`1.4
`1.4
`
`Container
`Type
`
`Vial
`Vial
`Vial
`Vial
`Vial
`Vial
`Ampoule
`Vial
`Ampoule
`Vial
`Vial
`Vial
`Vial
`Vial
`
`Vol. 43. No. 2 / March-April 1989
`
`71
`
`Opiant Exhibit 2323
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685, IPR2019-00688, IPR2019-00694
`Page 6
`
`

`

`TABLE V
`
`. Ch emica
`dies^
`
`Composition
`
`of Glasses—IS Stu-
`
`Downloaded from
`
`Chemical
`Components
`%
`
`IS-l
`
`LS-2
`
`I S -3
`
`Glass tt
`IS-4
`l'S-5
`
`LS-6
`"• -»
`10
`7
`7
`
`T
`
`1
`1
`
`IS-7
`
`7 "*
`10
`6
`7
`
`-i
`
`1
`2
`
`SO
`13
`4
`3
`
`-i~>
`
`10
`6
`7
`~\
`1
`i
`
`72
`10
`6
`7
`2
`1
`
`-t
`
`68
`10
`9
`8
`2
`1
`1
`
`1
`
`69
`10
`6
`6
`->
`1
`-)
`1
`
`3
`
`SiO^
`B:0,
`Na.O
`ALO,
`BaO
`CaO
`K,0
`Fe:03
`ZnO
`TiO,
`MgO
`J Omission of a value indicates that the material is not a component of
`the formulation; it may be present at trace levels.
`
`on April 13, 2020
`posed to aqueous media with a pH in the range of 4 to '
`8. Significantly greater levels of extractables w^ere ob(cid:173)
`served for all the borosilicate glass compositions when
`they were exposed to very alkaline media (pH > 9).
`Neither observation was surprising. Borosilicate con(cid:173)
`tainers are normally very resistant to attack by aque(cid:173)
`ous media except for very alkaline media (7).
`2. Si was one of the major extractables for all the glasses,
`regardless of the pH of the extraction medium. More(cid:173)
`over, as evidenced by the results in Tables XIV and
`XV. Si was the dominant extractable when the glasses
`were exposed to very alkaline conditions. It is notewor(cid:173)
`thy that Si was probably present in the form of silica
`and various silicates. In fact, the bands associated with
`SiO: and/or silicates (51) dominate the spectrum in
`Figure 1, which was typical of the infrared spectra
`observed for the extractables when the pH of the media
`was >9. Because of the potentially numerous silicon-
`containing compounds, no attempt was made to mea(cid:173)
`sure each species. Instead, the total amount of Si was
`measured and reported as SiOi.
`The aforementioned observations concerning SiOi
`are consistent with the knowledge of glass/media in(cid:173)
`teractions (7). In particular. SiO? is leached from a
`borosilicate glass via a dissolution mechanism, and for
`basic solutions the dissolution mechanism is the domi(cid:173)
`nant interaction between a borosilicate glass and an
`aqueous medium (7).
`The mechanism involving the dissolution of a boro(cid:173)
`silicate glass is complex, but the first step is probably
`the reaction of the glass with hydroxide ion (OH~)
`
`TABLE VII.
`
`omposition of
`Chemical C
`dies"
`
`Glasses—UK Stu-
`
`Chemical
`Components
`%
`
`LK-1
`
`UK-2
`
`LK-3
`
`Glass #
`UK-4
`
`LK-5
`
`LK-6
`
`IK-7
`
`75
`10
`6
`5
`2
`1
`
`74
`9
`7
`5
`2
`1
`1
`
`74
`9
`7
`5
`i
`1
`1
`
`SiO,
`B:03
`N a ,0
`AUO,
`BaO
`CaO
`K,0
`Fe:0,
`ZnO
`TiO,
`0.1
`MgO
`" Omission of a value indicates that the material is not a component of
`the formulation; it may be present at trace levels.
`
`67
`i:
`9
`6
`3
`1
`
`"1
`
`69
`11
`9
`6
`3
`0.5
`
`1
`
`1
`
`67
`11
`9
`5.7
`3.1
`1.5
`1
`1
`0.7
`
`71
`9
`6
`5
`~\
`1
`1
`1
`
`3
`
`experiments were performed. This preparation is similar
`to the procedures used in the USP Water Attack Test and
`the PDA protocol. Consequently, our studies did not in(cid:173)
`vestigate any of the effects of processing, even though
`washing treatments, drying conditions, methods of steril(cid:173)
`ization, and other processing steps can affect the level of
`extractables from a glass container (37, 39).
`
`Extraciable Studies
`
`Tables VIII-XV contain the results of the US and UK
`studies involving unbuffered extraction media. Besides
`data pertaining to each extractable concentration, these
`tables contain information for total extractables. pH mea(cid:173)
`surements, and results that have been normalized to the
`surface area of glass exposed to the extraction media.
`Tables XVI and XVII contain the results of some addi(cid:173)
`tional studies with US glass containers involving both
`buffered and unbuffered extraction media. Although
`there is a significant amount of information in these ta(cid:173)
`bles, some general conclusions can be made and these will
`be discussed below.
`
`Attack by OH" would be expected to occur more
`readily than attack by H^O at least for basic solutions
`where there is an appreciable concentration of hydrox(cid:173)
`ide ions. Subsequent reaction of the second, third, and
`
`rx
`
`s
`
`¿*
`
`4000
`
`2000
`
`WAVENUMBERS
`
`1200
`
`G00
`
`1. Only relatively low levels of extractables were leached
`from most of the glass containers when they were ex(cid:173)
`
`Figure 1—Infrared spectrum of residue from one of the extractable
`studies at pH =; 10.4.
`
`72
`
`Journal of Parenteral Science & Technology
`
`Opiant Exhibit 2323
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685, IPR2019-00688, IPR2019-00694
`Page 7
`
`

`

`Downloaded from
`
`on April 13, 2020
`
`TABLE Vili. Extractable Results-
`Measurement
`US-1
`
`= 4.10
`US Studies: pH =
`US-3
`US-2
`
`2.2
`
`4.1
`7.7
`+3.6
`
`33.0
`3.1
`0.54
`3.9
`0.9
`ND
`ND
`ND
`ND
`
`ND
`
`4.1
`4.8
`+0.7
`
`0.7
`0.61
`ND
`ND
`ND
`ND
`ND
`ND
`ND
`
`L'S-4
`
`ND
`
`4.1
`4.7
`+0.6
`
`0.3
`0.38
`0.14
`ND
`0.2
`ND
`ND
`ND
`ND
`
`Total extractables
`(mg in 100 mL)
`Initial pH
`Final pH
`pH change
`
`SÌO2 (ppm)
`Na (ppm)
`K (ppm)
`Al (ppm)
`Ba (ppm)
`Ca (ppm)
`Mg (ppm)
`Fe (ppm)
`Zn (ppm)
`
`Si02 (Mg/cm2)
`Na (fig/cm2)
`
`0.6
`
`4.1
`5.8
`+ 1.7
`
`0.5
`1.3
`0.16
`ND
`ND
`ND
`ND
`ND
`ND
`
`0.1
`0.32
`
`Normalized Data [Constant
`12.0
`0.2
`1.1
`0.22
`
`Surface Area]
`0.1
`0.14
`
`TABLE IX. Extractable
`Measurement
`
`
`
`Results—UK Studies: pH = = 4.19
`
`UK-1
`
`1.3
`
`UK-2
`
`2.3
`
`UK-3
`
`0.4
`
`UK-4
`
`0.2
`
`US-5
`
`0.1
`
`4.1
`4.6
`+0.5
`
`0.1
`0.38
`ND
`ND
`ND
`ND
`ND
`ND
`ND
`
`0.04
`0.14
`
`UK-5
`
`ND
`
`4.2
`4.5
`+0.3
`
`US-6
`
`ND
`
`4.1
`4.9
`+0.8
`
`0.3
`0.65
`ND
`ND
`ND
`ND
`ND
`ND
`ND
`
`0.1
`0.28
`
`UK-6
`
`0.6
`
`4.2
`4.6
`+0.4
`
`US-7
`
`ND
`
`4.1
`4.9
`+0.8
`
`0.2
`0.91
`0.29
`ND
`ND
`ND
`ND
`ND
`ND
`
`0.05
`0.23
`
`UK-7
`
`0.4
`
`4.2
`4.4
`+0.2
`
`Total extractables
`(mg in 100 mL)
`Initial pH
`Final pH
`pH change
`
`Si02 (ppm)
`Na (ppm)
`K (ppm)
`Al (ppm)
`Ba