In Re Gabapentin Patent Litigation

393 F. Supp. 2d 278, 2005 WL 2651284

Opinion

MEMORANDUM AND ORDER

LIFLAND, District Judge.

Plaintiffs Pfizer, Inc., Warner-Lambert Co., and Godeeke Aktiengesellschaft (collectively, “Warner-Lambert”) brought suit against Defendants Purepac Pharmaceutical Co. and Faulding Inc. (collectively, “Purepac”), Teva Pharmaceuticals USA, Inc., Teva Pharmaceutical Industries, Ltd. (collectively, “Teva”), Zenith Laboratories, Inc., Zenith Goldline Pharmaceuticals, Inc., and IVAX Corporation (collectively, “Ivax”), Apotex Corp., Apotex Inc., and TorPharm, Inc. (collectively, “Apotex”), and EON Labs Manufacturing, Inc. (“Eon”), alleging infringement of U.S. Patent No. 6,054,482, entitled “Lactam-Free Amino Acids” (“ ’482 patent”). 1 Defendants jointly move for summary judgment of noninfringement of the ’482 patent based on Warner-Lambert’s inability to meet its burden of proof. For the reasons set forth herein, Defendants’ Motion will be granted.

BACKGROUND

Neurontin® is sold by Warner-Lambert as an aid in the treatment of epileptic seizures and other cerebral disorders. The active ingredient in Neurontin® is a chemical compound called gabapentin. The chemical makeup of gabapentin allows the molecule, under certain conditions, to undergo an internal chemical reaction that results in a new compound called “gaba-pentin lactam.” Gabapentin lactam, which is more than twenty-five times as toxic as gabapentin, actually causes rather than prevents seizures. Given the serious risks posed by gabapentin lactam, Warner-Lambert scientists sought to minimize its formation. They did so by way of a process ultimately disclosed in the ’482 patent.

A. Development of Neurontin

Warner-Lambert scientists discovered that a stable formulation of gabapentin required careful limiting in two areas. First, it was critically important to ensure that the gabapentin itself was relatively pure. Second, the nature and amount of other additives in a gabapentin formulation affected stability. The former is pertinent to the instant motion.

When making a drug to market commercially, pharmaceutical companies generally prefer to use a “salt” of a drug in order to promote good stability and solubility, and to ensure that it is not irritating at the site of administration. A “salt” is formed as a result of the reaction between an acid and a base wherein a hydrogen ion from the acid is transferred to the base. Drugs that are “basic,” like gabapentin, form a salt with hydrochloric acid (HC1). Upon discovering that the hydrochloride salt of gabapentin turned to lactam more quickly than gabapentin itself, Warner-Lambert scientists turned their attention from using the hydrochloride salt to using “free” gabapentin.

When gabapentin is made by converting gabapentin hydrochloride (the acid form of gabapentin) to free gabapentin, any leftover acid corresponds to the mineral acid hydrochloric acid (HC1). Warner-Lambert maintains that its inventors, knowing that hydrochloric acid contained one chloride ion for each acid ion and that the only source of chloride ion in their process was hydrochloric acid, elected to measure the amount of chloride ion in their samples as a means of determining whether the amount of mineral acid was such that would allow for a stable sample. The inventors set a limit of 20 parts per million (“ppm”) by weight of chloride ion, reportedly believing that such a limit would guarantee that the amount of the mineral acid hydrochloric acid (HC1) is below 20 ppm and would allow Warner-Lambert to distinguish its gabapentin from unstable samples made in the past. 2

B. The’482 Patent

The ’482 patent includes eleven elaims-three independent claims (1,3,7) and eight dependent claims. Every claim in the ’482 patent includes the limitation that gaba-pentin have “less than 20 ppm” of an anion of a mineral acid or that an anion of a mineral acid does “not exceed 20 ppm.” An anion is a negatively charged ion. Mineral acids include nitric, sulfuric, and hydrochloric acids. An anion of hydrochloric acid, for example, is chloride.

Claims 1 through 6 of the ’482 patent are process claims. (’482 patent, col. 6,1. 55 to col. 8, 1. 28). Independent claim 1 concerns a process for making gabapentin. Claim 1 reads as follows:

A process for the preparation of a compound of Formula VII [gabapentin] ... comprising: ... (b) converting the acid addition salt of a compound of Formula VII by ion exchange to a compound of Formula VII containing less than 0.5% by weight of a compound of Formula VIII and less than 20 ppm of an anion of mineral acid.

(’482 patent, col. 6, 11. 55 — col. 7, 1. 20). Claim 2 depends from claim 1 and, accord ingly, includes the “less than 20 ppm” limitation. (Id. at col. 7, 11. 21-22). Claim 3, the next independent claim, teaches a process for making a pharmaceutical composition containing gabapentin:

A process for preparing stable and pure pharmaceutical compositions containing a compound of Formula (VII) [gabapen-tin] ... consisting of the steps of ... converting the acid additional salt of a compound of Formula VII by ion exchange to a compound of Formula VII containing less than 0.5% by weight of a compound of Formula VIII, wherein the proportion of the remaining anion of mineral acid does not exceed 20 ppm

(Id. at col. 7, 11. 23-57; col. 8, 11. 1-16). Claims 4, 5, and 6 depend from claim 3 and, therefore, include the “not exceed 20 ppm” limitation.

Claim 7 of the ’482 patent, the only independent product claim, covers a pharmaceutical composition containing gaba-pentin:

7. A stable and pure pharmaceutical composition in unit dry medicinal dosage form consisting essentially of:

(i) an active ingredient which is gaba-pentin in the free amino acid, crystalline anhydrous form containing less than 0.5% by weight of its corresponding lac-tam and less than 20 ppm of an anion of a mineral acid and

(ii) one or more pharmaceutically acceptable adjuvants that do not promote conversion of more than 0.2% by weight of the gabapentin to its corresponding lactam form when stored at 25 C and an atmospheric humidity of 50% for one year.

(Id. at col. 8, 11. 29-40 (emphasis added)). Claims 8-11 depend from claim 7 and, therefore, incorporate all the limitations of claim 7, including the “less than 20 ppm” limitation.

The written description of the ’482 patent states:

The active materials of formula (I) must be prepared as highly purified, nonderi-vatized free amino acids, for example, from the corresponding hydrochloride by ion exchange. The proportion of remaining hydrochloride admixture should thereby not exceed 20 ppm. The same also applies to other mineral acids.

(Id. at col. 5,11. 27-29).

C. ’482 Patent Prosecution History

The Godecke scientists applied for a German patent on their discovery in August 1989. They filed a corresponding United States patent application in August 1990. The grant of the ’482 patent, on April 25, 2000, came after a series of continuation United States patent applications and multiple reviews by the patent examiner.

None of the claims as originally filed included the “less than 20 ppm” or “not exceed 20 ppm” limitation. Claim 8 in the application became patent claim 1. Application claim 8 did not initially include the “less than 20 ppm” limitation. Warner-Lambert inserted the “less than 20 ppm” limitation into claim 8 following a rejection under Sections 102 and 103 based on the prior art Satzinger patent, U.S. Patent No. 4,024,175.

Application claim 10 became patent claim 3. Claim 10 initially specified that “the proportion of remaining mineral acid admixtures does not exceed 20 ppm.” The examiner rejected claim 10 as being unpat-entable due to prior art and for failing to satisfy the definiteness requirement for patentability. Among other things, the examiner believed that the language “the proportion of remaining mineral acid admixtures does not exceed 20 ppm” was ambiguous. To overcome the rejections, Warner-Lambert amended application claim 10 to state that “the proportion of remaining anion of a mineral acid does not exceed 20 ppm.”

Application claim 5 provided the foundation for patent claim 7. During prosecution, Warner-Lambert canceled claim 5 and replaced it with claim 21, and subsequently canceled claim 21 and replaced it with claim 24. Application claim 24 issued as patent claim 7. Pharmaceutical claim 5 referred to and incorporated compound claim 1. The examiner rejected claims 1 and 5 as being unpatentable due to the prior art. Warner-Lambert amended claims 1 and 5 in response to those rejections. When claim 21 replaced claim 5, it contained the “less than 20 ppm” limitation. And when claim 24 replaced claim 21, it too contained the “less than 20 ppm” limitation.

D. Defendants’ Abbreviated New Drug Applications

Apotex, Teva, and Ivax submitted Abbreviated New Drug Applications (“AN-DAs”) that do not set forth a minimum chloride value for the gabapentin and set an upper limit of 100 ppm. Test data submitted with their AND As demonstrate that their gabapentin formulations will meet the stability requirements of the ’482 patent claims. The biobatches used for the stability tests were made with a “low chloride” gabapentin-i.e., gabapentin that contains less than 20 ppm total chloride. The stability data submitted with the AN-DAs of Apotex, Teva, and Ivax was obtained from a product made with the same “low chloride” gabapentin.

Purepac and Eon amended their gaba-pentin ANDAS, setting a total chloride limit higher than 20 ppm in their gabapen-tin capsule and tablets. Purepac’s December 2000 amendments reflect that Purepac will only use gabapentin that contains a minimum of 28 ppm total chloride in its proposed product. Eon stated that it will only use gabapentin that contains a minimum of 30 ppm total chloride in its proposed product.

None of Defendants’ AND As specify quantities of chloride derived from mineral acid.

Defendants must use Teva’s gabapentin active pharmaceutical ingredient (“API”) in their products due to commitments to do so in their AND As. 3

E. Analytical Tests Performed on Ga-bapentin Samples

1. Teva’s Testing

Warner-Lambert maintains that Teva halted production of gabapentin shortly after issuance of the ’482 patent and implemented changes in its manufacturing process that resulted in an overall increase in the amount of total chloride ion in its gabapentin, but not in the amount of chloride ion derived from mineral acid. It is undisputed that Teva’s reformulated bulk gabapentin contains more than 20 ppm of total chloride anion (“high chloride” gaba-pentin). (Declaration of Peter J. Curtin (“Curtin Deck”), Ex. 4, Bartlett Initial Report at 45-46; Ex. 7, Dasgupta Report, ¶¶ 23-24; Ex. 3, Posner Rebuttal Report ¶¶ 90-92). Warner-Lambert’s expert, Dr. Paul A. Bartlett, identified potential sources for the residual chloride anion in Teva’s bulk gabapentin as gabapentin hydrochloride (an “acidic source”), tributyla-mine hydrochloride, and sodium chloride (a “non-acidic source”). Bartlett hypothesizes that the chloride values in Teva’s gabapentin largely reflect chloride from “non-acidic” sources and that the Teva ga-bapentin thus falls within the scope of the ’482 patent claims. (Curtin Decl., Ex. 4, Bartlett Initial Report at 15, 31-42; Ex. 5, Bartlett Dep. at 82:10-83.3; 161:12— 161:18).

Defendants conducted a variety of analytical testing on Teva’s gabapentin to measure the anion chloride and the cations sodium and tributylamine (“TBA”) therein. Dr. Purnendu K. Dasgupta analyzed twenty different lots of Teva’s gabapentin for chloride anion using a Dionex Ion Chroma-tograph ™ system, and conducted multiple analyses of each gabapentin lot. Dr. Das-gupta reported that the average chloride anion content for the twenty samples of Teva gabapentin ranged from 27.8 ppm to 82.5 ppm. (Curtin Decl., Ex. 7, Dasgupta Report ¶¶ 23-24, 44-45).

Dr. Ramon Barnes analyzed twenty different lots of Teva’s gabapentin for sodium using Inductively Coupled Plasma-Atomic Emission Spectrometry (“ICP-AES”), and conducted multiple analyses of each gaba-pentin lot. (Curtin Decl., Ex. 9, March 14, 2002 Barnes Report ¶¶ 20-22, 30). Dr. Barnes reports that the average sodium content for the samples of Teva’s gabapen-tin ranged from 0.62 ± 0.03 ppm to 6.7 ± 0.3 ppm.

Dr. Stephen MacDonald analyzed twenty lots of Teva’s bulk gabapentin for TBA using gas chromatography (“GC”), and conducted multiple GC analyses of each gabapentin lot. Most of the lots tested contained TBA levels below the quantifiable limit. The highest TBA content recorded for the twenty gabapentin samples averaged 2.55 ppm ± 0.55 ppm. (Curtin Decl., Ex. 10, March 14, 2002 MacDonald Report ¶ 7). Dr. Harold McNair also studied the levels of TBA in Teva’s bulk gabapentin, as measured by GC. Dr. McNair agrees with Dr. MacDonald that there are only trace amounts of TBA in the Teva gabapentin samples. (Curtin Deck, Ex. 11, March 14, 2002 McNair Report ¶ 13).

Warner-Lambert’s expert, Dr. Bartlett, agreed with the conclusions of Drs. MacDonald and McNair regarding the expected TBA content in Teva’s bulk gabapentin. Dr. Bartlett testified that he did not believe tributylamine hydrochloride would be a significant source of residual chloride anion in Teva’s gabapentin. (Curtin Deck, Ex. 5, Bartlett Dep. at 121:23-122:4).

Defendants argue that the foregoing results demonstrate that there can be no infringement because, even subtracting all the other sources of residual chloride anion identified by Dr. Bartlett, the minimum level of “acidic chloride” in Teva’s bulk gabapentin still exceeds both the 20 ppm limit disclosed in the ’482 Patent and the 22 ppm total chloride anion disclosed in the prior art. (Curtin Deck, Ex. 3, Table 1 to Expert Rebuttal Report of Dr. Gary Posner).

2. Warner-Lambert’s Testing

Warner-Lambert has not submitted any chloride or cation test data to refute Defendants’ test results. Warner-Lambert declined to replicate Defendants’ testing because “it probably wasn’t fruitful.” (Curtin Deck, Ex. 5, Bartlett Dep. at 256:13-17). Instead, Warner-Lambert measured the pH levels of Teva’s gabapen-tin to indirectly measure the level of “acidic chlorides” therein. (Curtin Deck, Ex. 12, Deposition of Martyn C. Davies, August 9, 2002 at 27:5 — 30:10).

The pH test measures the acidity of a solution. (Bartlett Deck ¶29). In the case of gabapentin, a solution of gabapen-tin is prepared by mixing gabapentin and water. (Id.). The solution is then measured by a pH meter. (Id.). The pH measurement for a sample of gabapentin containing high amounts of residual hydrochloric acid will be less than the pH measurement for a sample of gabapentin containing lesser amounts of residual hydrochloric acid. (Id.). Conversely, the pH is higher with less acid present in the sample (Id.).

The comparative pH method begins with the preparation of various gabapentin standard samples containing different known amounts of acid. (Bartlett Decl. ¶ 30). These standards are prepared from a sample of gabapentin that is virtually free of any residual hydrochloric acid (HC1). (Id.). The measured pH of the “unknown” gabapentin sample is then compared to that of the various standards with their known values of acid ion and chloride ion. (Id.). If the pH of the unknown sample measures higher than a particular standard, then that sample must contain less acid than the standard. (Id.).

Warner-Lambert contends that pH measurements reported by Teva for this “high chloride” gabapentin indicate that the acid ion content of Teva’s gabapentin had not increased as a result of Teva’s process changes. (Bartlett 1/17/03 Decl. ¶ 43). Thus, according to Warner-Lambert, the amount of acid in Teva’s new gabapentin remains low and falls within the 20 ppm limitation of the ’482 patent. That conclusion is based on the following: The pH measurement for a standard 18 ppm sample is 6.96. The pH measurements for five Teva samples tested are higher than 6.96, which means they have less acid content. (Bartlett 1/17/03 Decl. at ¶ 55).

One of Warner-Lambert’s experts, Dr. Martyn C. Davies, testified in his deposition as to the discrepancy between the high total chloride values and the less than 20 ppm chloride from acid values:

I show in my report ... the fact that all the pH values lie within .1, .2, ... pH unit relative to the Warner-Lambert sample. It’s extraordinary close pH values for this whole range of samples which repeatedly have different chloride levels. It’s clear ... that the chloride levels have nothing to do with acid.

(Lorenz Decl., Ex. 18 ¶ 59,11. 8-17).

According to Warner-Lambert, the comparative pH technique allows one to determine whether an unknown sample falls within the acid limitation of the ’482 patent, 4 (id.), and a comparison of the pH results of the Teva samples with standard gabapentin samples demonstrated that none of the Teva samples that Dr. Davies analyzed contained greater than 20 ppm chloride derived from acid-even though the total chloride ion content was greater than 20 ppm. (Lorenz Decl. ¶¶ 59-60). Thus, according to Warner-Lambert, Teva’s samples all fell within the ’482 patent’s 20 ppm limitation. (Bartlett 1/17/03 Decl. ¶ 61).

The pH tests were not designed to quantify the level of “acidic chlorides” in any particular sample, but were designed to compare the acidity levels of various gabapentin samples. Also, the pH measurements are subject to precision limitations. Dr. Bartlett testified that the limits of the precision of pH measurements are such that even the most carefully conducted pH measurements of gabapentin cannot detect differences in “acidic chloride” content of less than 10 ppm. (Curtin Decl., Ex. 5, Bartlett Dep. at 268:3-269:1; 378:17-374:7). He acknowledged that pH measurements could not necessarily distinguish between 20 ppm versus 18 ppm versus 22 ppm. (Id. at 266:19-24). He explained the limitations of the pH comparison method as follows:

If I have sample X, with a known amount of acid contaminant, and sample Y, with an unknown amount of acid contaminant, I believe that, with pH, you can tell the difference between them within a range of ten parts per million of chloride.

(Curtin Deck, Ex. 5, Bartlett Dep. at 269:25-270:6).

STANDARD OF REVIEW

Summary judgment is a procedural tool that obviates the need for trial by identifying and disposing of groundless claims and defenses. Celotex Corp. v. Catrett, 477 U.S. 317, 323-24, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986). Relief is warranted where “the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law.” Fed. R.Civ.P. 56(c). To resist, the adverse party must set forth specific facts that demonstrate the existence of a genuine issue for trial and may not rest on bare allegations or unsubstantiated defenses. Fed.R.Civ.P. 56(e); Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986).

The mere existence of some alleged factual dispute between the parties will not defeat an otherwise properly supported motion for summary judgment. Once the proponent discharges its Rule 56(c) duty, the burden shifts to the adverse party to show that material facts are genuinely controverted. Matsushita Electric Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 586, 106 S.Ct. 1348, 89 L.Ed.2d 538 (1986).

Materiality and genuineness are the touchstones of summary judgment law. A dispute is genuine only if the evidence is such that a reasonable fact-finder could find in favor of the nonmoving party. Anderson, 477 U.S. at 248, 106 S.Ct. 2505; Matsushita, 475 U.S. at 587, 106 S.Ct. 1348. To determine whether the proofs create a jury question, the Court must take into account the apposite evidentiary burden. Anderson, 477 U.S. at 254-55, 106 S.Ct. 2505. If the proofs presented would permit a jury applying the governing evidentiary standard to find for the adverse party, then a genuine factual dispute exists. Id. at 255, 106 S.Ct. 2505. Whether those disputed facts are material depends on the applicable substantive law. Id.

Because summary judgment involves a pretrial adjudication on the merits, the adverse party enjoys the benefit of various procedural protections. For example, the Court must view the evidence in the light most favorable to the adverse party and accord that party the benefit of all legitimate inferences. Matsushita, 475 U.S. at 587, 106 S.Ct. 1348. Moreover, the Court may not take credibility issues from the fact-finder. Anderson, 477 U.S. at 255, 106 S.Ct. 2505.

“Summary judgment is as appropriate in a patent case as it is in any other case.” C.R. Bard, Inc. v. Advanced Cardiovascular, Inc., 911 F.2d 670, 672 (Fed. Cir.1990). Where the material facts regarding the contents of the accused product are not in dispute, “the question of whether [the accused product] literally infringes the asserted claim of the ... patent turns on the interpretation of those claims,” which is purely a legal question for the court amenable to summary judgment. K-2 Corp. v. Salomon S.A., 191 F.3d 1356, 1362 (Fed.Cir.1999) (citing Athletic Alternatives, Inc. v. Prince Mfg., Inc., 73 F.3d 1573, 1578 (Fed.Cir.1996)).

DISCUSSION

A. ANDAs

“An infringement inquiry provoked by an ANDA filing ... is focused on the product that is likely to be sold following [Food and Drug Administration (“FDA”) ] approval. This determination is based on consideration of all the relevant evidence, including the ANDA filing, other materials submitted by the accused infringer to the FDA, and other evidence provided by the parties.” Abbott Labs. v. Torpharm, Inc., 300 F.3d 1367, 1373 (Fed.Cir.2002).

Warner-Lambert argues that the Apo-tex, Teva, and Ivax ANDAs speak to a level of chloride in the gabapentin anywhere from 0 ppm to 100 ppm. That range, of course, includes the infringing range of less than 20 ppm chloride. Warner-Lambert invites this Court to take the next step and assume that Defendants will sell gabapentin with the lowest possible chloride permitted by their ANDA and, in doing so, will infringe the ’482 patent.

The Court will not confine its analysis to Defendants’ ANDAs, which do not address the precise infringement issue before the Court. Rather, the Court will look to all the relevant evidence, including biobatch data results and additional testing of Teva’s gabapentin. See Abbott Labs., 300 F.3d at 1373. Here, Teva’s optimized manufacturing process for gabapentin is recorded in the Drug Master File that it filed with the FDA. The manufacturing process and testing conducted on samples made from that process are clearly relevant to the infringement inquiry.

B. Proof of Infringement

The claim construction dispute here is relatively straightforward. Warner-Lambert argues that the claim limitation “less than 20 ppm of an anion of a mineral acid” must be construed to mean less than 20 ppm of “acidic chlorides” (i.e., chloride anions derived from a mineral acid), rather than less than 20 ppm of total chloride anions regardless of their source. This distinction is particularly significant because Warner-Lambert’s expert has opined that potential sources for the residual chloride anion include non-acidic sources, such as sodium chloride. (Curtin Deck, Ex. 5 Bartlett Dep. at 82:10-83:3; 161:12-18; Ex. 4, January 31, 2002 Bartlett Initial Report at 15, 31-42, 46.). Defendants argue that the claim limitation must be construed to mean less than 20 ppm of total chloride anions regardless of their source. 5

Defendants move for summary judgment of noninfringement of the ’482 patent, arguing that the undisputed evidence of record demonstrates that Warner-Lambert, even under its own claim construction, cannot meet its burden to prove by a preponderance of the evidence that Defendants’ gabapentin products meet each and every limitation of the asserted ’482 patent claims, either literally or under the doctrine of equivalents.

1. Literal Infringement

Literal infringement is determined in a two-step process. First, a court must determine a claim’s acquired meaning and scope. Markman v. Westview Instruments, Inc., 52 F.3d 967, 976 (Fed.Cir.1995) (in banc), aff'd, 517 U.S. 370, 116 S.Ct. 1384, 134 L.Ed.2d 577 (1996). Second, the claim as construed must be compared to the accused product to ascertain whether it “reads” on the accused product. Southwall Technologies, Inc. v. Cardinal IG Co., 54 F.3d 1570, 1575 (Fed.Cir.1995). “To establish literal infringement, every limitation set forth in a claim must be found in an accused product, exactly.” Id. Where there is no dispute as to any relevant facts regarding the accused product, literal infringement is solely a matter of claim construction. Athletic Alternatives, Inc., 73 F.3d at 1578.

a. Parties ’ Arguments

Defendants have conducted testing on twenty lots of “high chloride” gabapentin to measure the level of chloride anions, sodium cations, and TBA in Teva’s gaba-pentin, which they claim demonstrates that the minimum level of “acidic chlorides” therein exceeds both the 20 ppm of the ’482 patent claims and the 22 ppm disclosed in prior art. Warner-Lambert has proffered no evidence of chloride testing or cation testing to refute Defendants’ test results. Instead, Warner-Lambert relies on pH measurements of Defendants’ gabapentin samples. As noted, the comparative pH testing involved comparing Teva’s gabapentin samples to standardized gabapentin samples containing known amounts of acid. There is no dispute that measuring the pH level of gabapentin does not quantify either its total chloride content or its “acidic chloride” content. Moreover, Warner-Lambert admits that isolated pH measurements cannot distinguish gabapentin falling within the scope of the ’482 patent claims {e.g., 18 ppm of “acidic chlorides”) from alleged prior art gabapentin (ie., 22 ppm of chloride anion). Defendants thus urge this Court to conclude that Warner-Lambert has proffered no competent evidence of infringement and therefore cannot meet its burden of proof in this case.

Warner-Lambert responds with several points. First, the ANDAs submitted by Defendants reflect gabapentin in the infringing range. Three of the five defendants-Apotex, Teva, and Ivax-have told the FDA in their respective ANDAs that them gabapentin will have no minimum level of chloride. That means those defendants are seeking to sell gabapentin that contains anywhere between 0 ppm and 100 ppm total chloride, which includes the infringing range. The biobatch data submitted to the FDA by these defendants further reveals use of “low chloride” gaba-pentin that indisputably contains less than 20 ppm chloride. The remaining two defendants-Purepac and Eon-have told the FDA that their gabapentin will have a minimum total chloride level of 28 ppm for Purepac and 30 ppm for Eon. The ANDAs say nothing about how much chloride derived from acid is in the gabapentin. Again, biobatch data reveals use of “low chloride” gabapentin that indisputably contains less than 20 ppm chloride. According to Warner-Lambert, there is no indication of using the “high chloride” ga-bapentin tested by Defendants’ experts. Thus, Warner-Lambert argues that Defendants have not demonstrated entitlement to summary judgment because the acidic chloride content of the gabapentin they intend to sell is not clear from their ANDAs.

Second, Warner-Lambert maintains that its testing of Defendants’ samples demonstrates that all the samples, including those containing total chloride as high as 77 ppm, actually contain less than 20 ppm chloride derived from mineral acid and thus fall within the claim limitation of the ’482 patent. Those test results, according to Warner-Lambert, create a clas sic material factual dispute that precludes summary judgment.

Finally, Warner-Lambert argues that Defendants infringe the ’482 patent even under Defendants’ claim construction-! e., 20 ppm chloride can be derived from any source, not just mineral acid. The argument goes that Apotex, Teva, and Ivax are free to sell gabapentin containing the lowest chloride level permitted by their AN-DAs, 0 ppm chloride, and products of Purepac and Eon, with chloride values of 28 ppm and 30 ppm, infringe the “less than 20 ppm chloride limitation” under the doctrine of equivalents.

Defendants reply that Warner-Lambert contravenes black-letter law because an infringement analysis demands comparison of the accused product with the language of the patent claims and not a product-to-product comparison. Zenith Labs., Inc. v. Bristol-Myers Squibb Co., 19 F.3d 1418, 1423 (Fed.Cir.1994) (“[I]t is error for a court to compare in its infringement analysis the accused product or process with the patentee’s commercial embodiment or other version of the product or process; the only proper comparison is with the claims of the patent.”). The argument goes that Warner-Lambert, through its comparative pH analysis, has performed a forbidden product-to-product comparison and offers the results as proof of infringement of the asserted claims.

In addition, Defendants point out that the ANDAs cited by Warner-Lambert are silent regarding the “acidic chloride” content of the accused products and thus do not address the relevant issue under Warner-Lambert’s proposed claim construction. Furthermore, the biobatches are irrelevant to the infringement inquiry insofar as they were all manufactured using gabapentin produced prior to the issuance of the ’482 patent. Teva’s optimized manufacturing process is recorded in the 2001 Drug Master File that it filed with the FDA. Teva’s process revisions substantially increased the level of residual chloride anions in Teva’s gabapentin, and the Teva gabapentin made according to that process contains more than 20 ppm of total chlorides. Teva must follow this optimized manufacturing process for all gabapentin that it manufactures for sale in the United States and, in turn, Defendants must use Teva’s gabapentin API in their products due to the commitments to do so in their ANDAs.

Defendants add that the asserted claims are written in terms of the amount of residual chloride anions in the gabapentin“less than 20 ppm of an anion of a mineral acid.” They are not written in terms of acidity, and there is no mention of pH measurements or pH levels in the asserted claims. Accordingly, Defendants argue that pH measurements cannot substitute for a measurement of infringement in terms of the specified claims. That would be tantamount to allowing Warner-Lambert to retreat from the plain language of the claims. Thus, according to Defendants, Warner-Lambert’s test results, which measure infringement in terms of comparative pH rather than residual chloride anions, cannot prove infringement of the ’482 patent claims. See Abbott Labs., 300 F.3d at 1376, Zenith Labs., 19 F.3d at 1423.

b. Analysis

As a threshold matter, the Court rejects Warner-Lambert’s invitation to assume that certain Defendants will market gaba-pentin with as low as 0 ppm chloride. It seems clear that Teva must manufacture its gabapentin consistent with the optimized process set forth in the Drug Master File, which yields a “high chloride” gaba-pentin containing more than 20 ppm of residual chloride anions. In any event, Teva represents that it must comply with the process set forth in the Drug Master File. The Court accepts and relies on that representation.

As to Warner-Lambert’s testing, the Court will not outright reject those test results as an impermissible product-to-product comparison. The comparative pH method assesses the acidity levels in the accused product. This method, by definition, involves a comparison of an accused product against standardized samples 6 with known quantities of mineral acid. The results from that comparison indicate a certain level of acidity, which is in turn compared against the literal claim terms. Thus, this method is not a direct product-to-product comparison. Whether that comparison suggests infringement in this case is of course another matter, to which the Court now turns.

The asserted claims of the ’482 patent all include the limitation that the gabapen-tin must contain less than “20 ppm” of “an anion of a mineral acid.” There is no dispute that the gabapentin manufactured by Teva since the ’482 patent issued-and that disclosed in the Drug Master File submitted to the FDA-contains more than 20 ppm of residual chloride anion. (Ex. 4, Bartlett Initial Report at 43-45; W-L Opp. Br. at Ex. B). While it is possible to measure the total chloride content (chloride anions) in gabapentin, there is no known analytical method to identify the cation with which a particular chloride anion had been associated. Measuring the level of chloride anion measures total chlorides regardless of their source. (Curtin Deck, Ex. 5, Bartlett Dep. at 98:1-18). It follows that measuring the total chloride levels of Teva’s gabapentin will not reveal whether that gabapentin falls within the scope of the asserted claims, as construed by Warner-Lambert.

The question then is whether Warner-Lambert can meet its burden to prove infringement of the asserted claims in terms of residual “acidic chlorides.” Warner-Lambert must be able to prove by a preponderance of the evidence that Defendants’ gabapentin products meet each and every limitation of the asserted ’482 patent claims either literally or under the doctrine of equivalents. Glaxo, Inc. v. Novopharm, Ltd., 110 F.3d 1562, 1565, 1570 (Fed.Cir.1997).

Warner-Lambert’s analytical expert, Dr. Davies, measured directly the acid content of Teva’s gabapentin samples. His test results contradict Defendants’ test results in a significant way. Dr. Davies compared Teva samples to a series of standard gabapentin samples containing different, but known, amounts of acid. (Bartlett 1/17/03 Decl. ¶ 48). Dr. Davies prepared his samples by starting with a baseline sample from Warner-Lambert that contained no detectable chloride. He then completed the preparation of his standards by adding varying known amounts of acid to this baseline sample. (Id.). Dr. Davies took the pH measurements of his standardized samples (Id. ¶ 49), which revealed that the pH of the samples changed inversely with the addition of mineral acid. Dr. Davies also added known amounts of acid to various Teva samples (Id. ¶ 50), all of which showed the identical pH decrease.

As to Teva’s “high chloride” gabapentin, all of the samples that Dr. Davies tested are stated to contain less than 20 ppm of mineral acid. (Id. ¶ 55). Each one contains less mineral acid than the standard sample containing 18 ppm hydrochloric acid. (Id.). This is demonstrated by the fact that the pH measurement for the standard 18 ppm samples is 6.96. The pH measurements for the five Teva samples are higher than 6.96, which, according to Dr. Davies, means they have less acid content. (Id.). A higher pH measurement means less acid. These test results suggest that the amounts of chloride in Defendants’ gabapentin are not indicative of the amount of acid and, instead, are in the form of some other non-acidic chloride salt.

Dr. Davies explained his conclusions as follows:

When I look at the solutions [of Teva’s gabapentin and of the standardized ga-bapentin sample], they are virtually identical in terms of pH ... What does that tell me? I know that ... if you had different amounts of acid, it would change the pH of these solutions. I know that ... if the chloride that is reported for Teva’s sample was related to acid, you would have a substantial shift on the pH as registered for the ... solution. The fact that they are all so tight and close clearly indicates these samples contain the same amount, less than 20 parts per million.

(Lorenz Deck, Ex. 18 at 142, 1. 10 — 143, 1. 5).

If the analysis were to end here, the Court would find that Dr. Davies’ comparative pH testing results created a genuine issue of material fact requiring denial of Defendants’ Motion; ie., a dispute between experts involving the relative merits of different scientific tests addressing infringement. However, Dr. Bartlett, Warner-Lambert’s expert, has testified to the limited precision of pH measurement which requires a different result.

Warner-Lambert maintains that Dr. Davies did not need to further determine whether the Teva samples actually contained 20 ppm or 19 ppm or some other precise amount. While that may be true, the Court feels that the appropriate question is whether the limited precision of measurement in the pH method testified to by Dr. Bartlett means that the same pH reading for the two samples would allow for a 10 ppm variation between them in acid content? In other words, after putting 18 ppm acid into the standard sample, and getting the exact same pH reading from the “unknown” sample, could the unknown sample still have between 13 and 23 ppm of acid? The Court’s review of the record causes it to answer that critical question in the affirmative.

Evidence of infringement via a comparative pH measurement does not quantify total chlorides or “acidic chlorides,” and pH measurements cannot distinguish between different levels of acidity reflected in a difference of 10 ppm of “acidic chlorides.” (See Curtin Decl., Ex. 5, Bartlett Dep. at 266:19-267:3; 268:3-269:1; 269:17-270:10; 313:20-314:10; 373:17-374:7.). As noted, Dr. Bartlett explained the limitations of the comparative pH method as follows:

If I have sample X, with a known amount of acid contaminant, and sample Y, with an unknown amount of acid contaminant, I believe that, with pH, you can tell the difference between them within a range of ten parts per million of chloride.

(8/27/02 Bartlett Dep. at 269:25-270:6). The Court understands that testimony to mean that the comparative pH method cannot detect differences in the acidity of gabapentin that would be reflected in changes in “acidic chloride” content of less than 10 ppm.

Given that admitted limitation in the comparative pH method, the question is whether that is fatal to Warner-Lam bert’s ability to prove infringement. Teva’s gabapentin API contains more than 20 ppm total chloride anion, but, according to Warner-Lambert, contains some as “acidic chlorides” and some not. Having drafted its claim language in terms of quantitative measurement of residual chloride anions, Warner-Lambert must prove infringement in quantitative terms. See, e.g., Abbott Labs. v. Torpharm, Inc., 300 F.3d at 1376; Zenith Labs., Inc. v. Bristol-Myers Squibb Co., 19 F.3d 1418, 1423.

In light of the admitted limitations of the pH method, the Court concludes that Warner-Lambert cannot satisfy its burden to prove that an accused product falls with the literal scope of the 20 ppm limitation. The pH levels are simply not precise enough for determining literal infringement of the claim in terms of acidic chloride content. In other words, pH readings for gabapentin samples containing 18 ppm of “acidic chlorides” will be indistinguishable from those for gabapen-tin samples containing 23 ppm or more.

The only remaining question is whether pH test results indicating acidic chloride content falling in or around the 20 ppm limitation, allowing for the limited precision of that measurement, might be sufficient to prove infringement under the doctrine of equivalents.

2. Infringement Under the Doctrine of Equivalents

Defendants argue that Warner-Lambert can neither assert nor demonstrate infringement under the doctrine of equivalents with respect to the 20 ppm limitation of the ’482 patent. The Court has considered the arguments offered for and against Defendants’ position, as amplified at oral argument, and concludes that Warner-Lambert is not entitled to assert a range of equivalents.

The “doctrine of equivalents” allows a patentee to claim insubstantial alterations not expressly captured in drafting of the original patent claim. Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki, 535 U.S. 722, 733, 122 S.Ct. 1831, 152 L.Ed.2d 944 (2002). The underlying rationale of the doctrine is that language may not capture the true essence of an invention. Id. at 734, 122 S.Ct. 1831.

There are two major limitations on the doctrine of equivalents. A patentee cannot recapture subject matter surrendered during prosecution to obtain patentability. Warner-Jenkinson Co. v. Hilton Davis Chem. Co., 520 U.S. 17, 30, 117 S.Ct. 1040, 137 L.Ed.2d 146 (1997). Nor can a patentee claim that which could not have been patented based on prior art. Marquip, Inc. v. Fosber Am. Inc., 198 F.3d 1363, 1367 (Fed.Cir.1999). Also, “[i]f a theory of equivalence would vitiate a claim limitation, ... there can be no infringement under the doctrine of equivalents.” Tronzo v. Biomet, Inc., 156 F.3d 1154, 1160 (Fed.Cir.1998).

In Festo, the Supreme Court held that a patentee bears the burden of proving that a narrowing amendment did not surrender the alleged equivalent in question by showing that at the time of the amendment one skilled in the art could not reasonably be expected to have drafted a claim that would have literally encompassed the alleged equivalent. Id. Examples where the patentee might satisfy its burden to establish that estoppel does not bar a claim of equivalence include: “The equivalent may have been unforeseeable at the time of the application; the rationale underlying the amendment may bear not more than a tangential relation to the equivalent in question; or there may be some other reason suggesting that the patentee could not reasonably be expected to have de scribed the insubstantial substitute in question.” Id.

Defendants make three main arguments in support of limiting the range of equivalents that may be asserted by Warner-Lambert. First, the ’482 patent claims specifically exclude products with more than 20 ppm of an anion of a mineral acid and a range of equivalents that exceeds that numerical range vitiates the claim.

Second, prosecution history estoppel arises because Warner-Lambert surrendered coverage through patentability arguments made in connection with amendments to claim elements. For instance, Warner-Lambert distinguished the ’482 patent’s claims from the ’175 Satzinger patent by asserting that compounds made by the Satzinger process would have a chloride content of 200 ppm, while the claimed compound needed to have a chloride content below 20 ppm. (Pros. Hist.000704-07, 733 — 34). 7 It also asserted that a chloride level of “less than 20 ppm” was critical to stability and that the prior art did not recognize the importance of a low chloride content. (Pros.Hist.000734, 744-45, 765, 774, 784, 796, 811, 832, 877, 893, 904, 906, 910). Warner-Lambert similarly distinguished claims of the ’482 patent from U.S. Patent No. 4,894,476 (“the Butler patent”). The ’476 Butler patent discloses, among other things, processes for the small- and large-scale production of gabapentin. In particular, Example 3, which concerns a large-scale process, describes gabapentin with a chloride content of 22 ppm, which is very near the ’482 patent’s “less than 20 ppm” and “not exceed 20 ppm” limitations. The Patent Examiner repeatedly rejected Warner-Lambert’s claims under 35 U.S.C. § 103 in view of the Butler ’476 patent, noting that it would be obvious to reduce the anion content of the gabapentin even further. (12/9/99 Second Submission After Final Rejection at 2-5, Ex. 11 to Goldsmith Decl. 4/30/96 Response to Office Action at 3-4). In response to those objections, Warner-Lambert stated that “[t]he anhydrous GA-BAPENTIN of Example 3 of Butler has 22 ppm of chloride anion. This is outside the claim limitation of the present invention.” (12/9/99 Second Submission at 5). Based on that limiting argument made during prosecution, Defendants urge the Court to conclude that the range of equivalents in this case is limited to at least 22 ppm of chloride anion.

Third, Defendants argue that Warner-Lambert has not carried its burden to demonstrate that the range of equivalents it seeks would not ensnare prior art disclosing a 22 ppm chloride limit. See Wilson Sporting Goods Co. v. David Geoffrey & Assocs., 904 F.2d 677, 685 (Fed.Cir.1990), overruled in part on other grounds, Cardinal Chem. Co. v. Morton Int’l, 508 U.S. 83, 113 S.Ct. 1967, 124 L.Ed.2d 1 (1993).

Warner-Lambert responds that it never maintained that the ’482 patent invention was patentable over the Butler patent based on the distinction between the claimed invention of 20 ppm chloride ion and the gabapentin sample reported in the Butler patent having 22 ppm chloride ion. Instead, Warner-Lambert maintains that the inventors essentially conceded that Butler described a gabapentin sample that had a chloride level within the scope of the ’482 patent claims, (Pros.Hist.000796), but distinguished Butler on the basis that it failed to teach anything about the delete rious effect of mineral acid or of certain adjuvants on the formation of gabapentin lactam. Warner-Lambert adds that this claim limitation (in somewhat different language) was initially added 20 months after the ’482 patent was filed and the actual language “less than 20 ppm of an anion of a mineral acid” was added 22 months later. This language was not added in response to a rejection based on Butler, which came another 20 months after the language was added.

The Court recognizes that the key difference between the ’482 patent and Butler is recognition of a stability problem and not necessarily the precise chloride limitation disclosed therein. Warner-Lambert might have obtained patentability, even asserting the same chloride limitation as that disclosed in Butler. For that reason the Court is not inclined to limit the range of equivalents based on Butler alone.

That said, the Court concludes that Warner-Lambert is not entitled to assert a range of equivalents with respect to the 20 ppm limitation. First, allowing Warner-Lambert to claim a range of equivalents higher than that allowed by the 20 ppm limitation arguably acts to vitiate the claim limitation. Since the claim is clearly directed to a quantity below a defined level, allowing a quantity above that defined level to be an equivalent deprives the “less-than” aspect of the claim of any meaning. Cf. Moore U.S.A., Inc. v. Standard Register Co., 229 F.3d 1091, 1106 (Fed.Cir.2000). However, the Court is persuaded that argument-based and amendment-based es-toppel both clearly apply.

As to argument-based estoppel, Warner-Lambert did argue that the invention should be distinguished from Butler, at least in part, on the basis of the 22 ppm limitation: “The anhydrous GABAPEN-TIN of Example 3 of Butler has 22 parts per million of chloride anion. This is outside the claim limitation of the present invention.” (Pros.Hist.000893).

As to amendment-based estoppel, Warner-Lambert concedes, as it must, that the limitation “less than 20 ppm of an anion of a mineral acid” was added in the face of a rejection based on its prior Satzinger patent. This amendment was clearly made for a reason related to patentability, and thus estoppel applies. The question then becomes what subject matter the amendment surrendered.

Warner-Lambert bears the burden of showing that the amendment did not surrender the particular equivalent it now seeks to claim. Warner-Lambert contends that there is a viable question as to whether it is entitled to a range of equivalents because 20 is not a “critical” number. (11/19/04 Tr. at 114). Rather, the 20 ppm limitation was put in to go as far down within a measurable range from prior art like Satzinger, claiming 200 ppm. At oral argument, Warner-Lambert urged that it could not reasonably be expected to have described the insubstantial substitute in question because an infringer is technically always able to get around a numerical claim limitation by going slightly above or below the claimed range.

Warner-Lambert made a narrowing amendment during prosecution in adding the limitation “less than 20 ppm of an anion of a mineral acid” and the amendment was made for reasons related to pat-entability. Warner-Lambert has failed to meet its burden of showing that it did not surrender chloride levels in excess of the “less than 20 ppm” limitation. It falls well short of its burden of showing that at the time this amendment was made, one of skill in the art could not reasonably have been expected to draft a claim that would encompass the alleged equivalent. To the contrary, one skilled in the art at the time of the amendment clearly could have reasonably been expected to draft a claim that would encompass higher chloride levels. Warner-Lambert’s own prior art supports that idea. The Satzinger patent disclosed up to 200 ppm chloride, the Hartenstein patent (U.S. Patent No. 4,152,326) disclosed 2000 ppm chloride, and the Butler patent disclosed 22 ppm chloride. Warner-Lambert clearly knew about heightened chloride levels at the time of amendment.

For the foregoing reasons, the Court concludes that Warner-Lambert is not entitled to assert a range of equivalents for the 20 ppm claim limitation of the ’482 patent.

CONCLUSION

For the reasons stated herein, Defendants are entitled to summary judgment of noninfringement based on Warner-Lambert’s inability to meet its burden to prove infringement, both literally and under the doctrine of equivalents.

Accordingly, IT IS on this 22nd day of August 2005,

ORDERED that the Motion of Defendants for summary judgment of nonin-fringement of U.S. Patent No. 6,054,482 based on Warner Lambert’s inability to meet its burden of proof is granted.

1 . Warner-Lambert filed separate patent infringement actions against multiple generic drug manufacturers. The Judicial Panel on Multidistrict Litigation directed that all such actions be consolidated before this Court for coordinated pretrial proceedings. This Motion pertains to the "first-wave” defendants. There are also "second-wave” and "third-wave” defendants, sued by Warner-Lambert after close of discovery for first-wave defendants.

2 . Defendants dispute and/or challenge the background information provided by Warner-Lambert as irrelevant to the key issue-the adequacy of Warner-Lambert's proffered evidence of infringement. Nonetheless, this background information is set forth because the Court believes that it will help in understanding Warner-Lambert's position herein.

3 . Apotex's ANDA initially designated co-defendant Teva as its sole supplier of gabapen-tin. Apotex has since added to its ANDA Hikal and Zambón as alternate suppliers of gabapentin. Gabapentin supplied by Hikal and Zambón is the focus of separate summary judgment motions.

4 . Alternative methods of measurement are described in detail in this Court's Memorandum and Order denying Defendants’ Joint Motion for summary judgment invalidating the claims of the '482 Patent as Indefinite under 35 U.S.C. § 112, ¶ 2.

5 . Warner-Lambert argued in its brief that the Court should hold a Markman hearing. At oral argument on the pending summary judgment motions, all parties, including Warner-Lambert, agreed that a separate Markman hearing was not necessary.

6 . The fact that those standardized samples happen to be Warner-Lambert products is irrelevant.

7 . The prosecution history cited to is that submitted in connection with Motion of Purepac Pharmaceutical Co. and Faulding Inc. for Summary Judgment of Noninfringement pertaining to the '482 patent's “20 ppm” limitation. See January 15, 2001 Declaration of Steven M. Amundson, Ex. B.