Pfizer Canada Inc. v. Canada (Health)

Pfizer Canada Inc. v. Canada (Health)
Court (s) Database
Federal Court Decisions
Date
2007-10-05
Neutral citation
2007 FC 898
File numbers
T-507-05
Notes
Digest
Decision Content
Date: 20071005
Docket: T-507-05
Citation: 2007 FC 898
Ottawa, Ontario, October 5, 2007
PRESENT: The Honourable Madam Justice Snider
BETWEEN:
PFIZER CANADA INC.
and WARNER-LAMBERT COMPANY, LLC
Applicants
and
THE MINISTER OF HEALTH
and RANBAXY LABORATORIES LIMITED
Respondents
Restriction on publication:
“These are the public version of sealed reasons, dated September 11, 2007, pursuant to the Protective Order dated April 7, 2005.”
REASONS FOR ORDER AND ORDER
1. Introduction
[1] Atorvastatin calcium is the active ingredient in LIPITOR, an anti-cholesterol drug marketed in Canada by the Applicants (collectively referred to as Pfizer or the Applicants). Canadian Patent No. 2,220,018 (the 018 Patent) claims and protects particular and novel crystalline forms of atorvastatin calcium called Form I, Form II, and Form IV. Canadian Patent No. 2,220,455 (the 455 Patent) is also listed on the Patent Register in respect of LIPITOR; it claims a novel process for making atorvastatin calcium in amorphous form from one of the crystalline forms of the 018 patent.
[2] Ranbaxy Laboratories Limited (Ranbaxy) manufactures Ran-Atorvastatin, which contains the amorphous form of atorvastatin calcium, in India. Pursuant to the relevant Patented Medicines (Notice of Compliance) Regulations, S.O.R./93-133 (NOC Regulations), Ranbaxy has applied to the Minister of Health (the Minister) for approval to sell its product into Canada. As required by the NOC Regulations, Ranbaxy served a Notice of Allegation (NOA) dated January 31, 2005, addressed to Pfizer Canada Inc., wherein Ranbaxy alleged that the six patents then listed on the Patent Register in association with Pfizer’s atorvastatin calcium products would either not be infringed by the manufacture and sale of Ran-Atorvastatin, were invalid, or both. This particular hearing deals with only two of those patents – the 018 Patent and the 455 Patent.
[3] Pfizer does not contest Ranbaxy’s factual assertion that Ran-Atorvastatin, the product Ranbaxy seeks to import into and sell in Canada, contains the amorphous form of atorvastatin calcium, and none of the patented crystalline forms. Rather, Pfizer contends that Ranbaxy uses one of the patented crystalline forms as an intermediate in the process to make the amorphous material contained in Ran-Atorvastatin, and that Ranbaxy infringes the process described in the 455 Patent. Meanwhile, Ranbaxy asserts that it infringes neither of the relevant patents and that the 455 Patent is invalid for insufficiency.
[4] For the reasons that follow, I have determined that the application will be allowed with respect to the 018 Patent but dismissed with respect to the 455 Patent.
2. Background
2.1 History of this Proceeding
[5] This is the second hearing in respect of the January 31, 2005 NOA. As this is a rather unusual occurrence, some explanation of how this second hearing came about may provide helpful background to the reader.
[6] As noted, the NOA referred to six patents. Initially, Pfizer contested each of Ranbaxy’s allegations for all six patents. Pfizer subsequently obtained leave to discontinue the application with respect to four of the six patents, resulting from the second disclosure of Ranbaxy. The application in respect of the two remaining patents (the 768 Patent and the 546 Patent) was heard in January 2007. By order dated January 25, 2007 (Pfizer Canada Inc. v. Canada (Minister of Health), 2007 F.C. 91), Justice von Finkenstein dismissed Pfizer’s application in respect of the 546 Patent. While Pfizer’s application in respect of the 768 Patent was allowed, that patent has since expired (in May 2007). Pfizer’s appeal of the January 25, 2007 order on the 546 Patent was heard on May 22 and 23, 2007. No order on that appeal has yet been issued.
[7] In the context of another NOC proceeding on patents related to LIPITOR, Pfizer became aware of additional facts about Ranbaxy’s manufacture of Ran-Atorvastatin that, in Pfizer’s view, raised further issues related to the 018 and 455 Patents. Pfizer sought and was granted leave to reinstate this proceeding for these two patents. (Pfizer Canada Inc. v. Canada (Minister of Health), (4 December 2006), Ottawa, T-507-5 (Fed. Proth.), aff’d 2007 FC 205, aff’d 2007 FCA 244). Accordingly, this hearing proceeded in relation to the 018 and 455 Patent.
2.2 Ranbaxy’s Process
[8] Ranbaxy makes two atorvastatin calcium intermediate products in the manufacture of its atorvastatin calcium API: Atorvastatin Calcium Crude (ATV-2) and a further purified intermediate, Atorvastatin Calcium (ATV-2P). In addition, ATV-2 is made using a seed (Atorvastatin Calcium Seed). Seeds are agents used to influence nucleation and induce crystallization. ATV-2, ATV-2P and the Atorvastatin Calcium Seed (together, referred to as the Ranbaxy intermediates) are all crystalline forms of atorvastatin calcium.
[9] The active substance or API produced by the Ranbaxy process is called Atorvastatin Calcium (Amorphous) (ATV-3). ATV-3 is amorphous atorvastatin calcium and is named Ran-Atorvastatin by Ranbaxy.
[10] The experts agree that Ranbaxy’s process for converting ATV-2P into ATV-3 involves dissolving ATV-2P in [compound A], a non-hydroxylic solvent, and then adding [compound B] to precipitate the ATV-3 from the solution. [Compound B] is a hydrocarbon solvent and is non-polar. Atorvastatin is a salt, and therefore polar. Polar substances are insoluble in non-polar solvents, and since the amount of [compound B] is far greater than [compound A], the solution is rapidly rendered insoluble and the ATV-3 precipitates to the bottom of the production vessel (Application Record of the Applicants [A.R.], Vol. 12, Tab 18, p. 3507).
3. Issues
[11] There are two sets of issues to be addressed in this proceeding – one set for each of the patents.
[12] With respect to the 018 Patent, the issues are as follows:
1. What is the proper construction of Claims 1 to 9 of the 018 Patent? Of specific interest, is crystalline Form I atorvastatin calcium uniquely identified by XRPD peaks and 13C NMR data described in the claims and is the substance claimed by Claims 1 to 9 a hydrate?
2. Is Ranbaxy’s NOA inadequate because it does not allege non-infringement: on the basis that the product is made in India; because it only contains a bald assertion in respect of claims 6 to 9; or because it does not clearly state that it is using the anhydrous form of crystalline atorvastatin calcium rather than the hydrate form?
3. Has Pfizer met its burden of satisfying this Court that Ranbaxy’s allegation of non-infringement of claims 1 to 9 of the 018 Patent is not justified? Subsidiary to this issue are the following:
(a) Is the matching of the X-ray powder diffraction (XRPD) peaks of the patented Form I crystalline atorvastatin calcium with the three intermediate forms of the atorvastatin calcium used by Ranbaxy in its manufacture of Ran-Atorvastatin, together with other testing carried out by Pfizer’s experts, sufficient to demonstrate that the three intermediate forms are covered by Claims 1 to 5 of the 018 Patent?
(b) By not providing evidence of the Carbon-13 Nuclear Magnetic Resonance (13C NMR) data of the patented Form I atorvastatin calcium data for its intermediates, has Ranbaxy failed to allege any facts to support its allegation of non-infringement of Claims 6 to 9 of the 018 Patent?
(c) If claims 1 to 9 are limited to hydrates, does Ranbaxy use the hydrate as an intermediate in the making of Ran-Atorvastatin?
(d) Does the use of the patented crystalline Form I atorvastatin calcium in India as an intermediate constitute infringement of the 018 Patent under Canadian patent laws?
[13] With respect to the 455 Patent, the relevant issues are as follows:
1. What is the proper construction of Claims 75 to 110 of the 455 Patent? Of particular relevance, are the claims to be construed as covering formation of amorphous atorvastatin calcium through techniques that include both evaporation and precipitation or as limited to evaporation?
2. Is Ranbaxy’s NOA inadequate because it does not raise the claim construction as an allegation?
3. Has Ranbaxy led sufficient evidence to rebut the presumption of validity on the basis of insufficiency and has Pfizer, in turn, failed to meet its burden of showing that the allegation of invalidity is not justified?
4. Witnesses
[14] Each of Pfizer and Ranbaxy provided affidavit evidence from a number of witnesses whose evidence addressed both technical and factual matters. The qualifications of the four witnesses put forward as experts were not disputed. In summary form, the following affiants were most relevant to the issues before me.
Pfizer’s Witnesses
Expertise/Background
Subject-Matter of Evidence
Dr. Allan S. Myerson
Provost & Senior Vice President and Philip Danforth Armour Professor of Engineering at the Illinois Institute of Technology; crystal form expert
Patents, Ranbaxy’s process and intermediates
Dr. Nair Rodríguez-Hornedo
Associate Professor of Pharmaceutical Sciences at the College of Pharmacy at the University of Ann Arbor, Michigan; crystal form expert
Patents, Ranbaxy’s process and intermediates
Mr. Rex Shipplett
Senior Scientist with SSCI Inc.
Fact witness regarding the sufficiency of the 455 Patent
Ranbaxy’s Witnesses
Expertise/Background
Subject-Matter of Evidence
Dr. Mark D. Hollingsworth
Associate Professor in the Department of Chemistry at Kansas State University; crystal form expert
Patents, Ranbaxy’s process and intermediates
Dr. Ian M. Cunningham
Chairman for Dynamic Extractions Ltd. from 2002-2004, and was a Senior Vice-President of Chemical Development at GlaxoSmithKline from 2001-2002; pharmaceutical expert
455 Patent, Ranbaxy’s process
Mr. Jay R. Deshmukh
Lawyer and Senior Vice President, Intellectual Property for Ranbaxy
Fact witness regarding Ranbaxy’s process
5. The 018 Patent
5.1 What is the proper construction of the 018 Patent?
[15] As taught by the jurisprudence, my first task is to undertake a “purposive construction” of the claims in issue. There is no disagreement and thus no need to set out an exhaustive list of the well-established principles of claims construction (see, principally, Free World Trust v. Electro Sante Inc., [2000] 2 S.C.R. 1024, 9 C.P.R. (4th) 168, and Whirlpool Corp. v. Camco Inc., [2000] 2 S.C.R. 1067, 9 C.P.R. (4th) 129). In sum:
The key to purposive construction is therefore the identification by the court, with the assistance of the skilled reader, of the particular words or phrases in the claims that describe what the inventor considered to be the "essential" elements of his invention (Whirlpool, above at para. 45).
[16] The first patent in issue is the 018 Patent. In the 018 Patent, the patentees describe the need to produce atorvastatin in a pure and crystalline form to enable formulations to meet exacting pharmaceutical requirements and specifications (A.R., Vol. 1, Tab 4, p. 82). The 018 Patent discloses and claims novel crystalline forms of atorvastatin calcium called Form I, Form II and Form IV. As described in the patent, all of the forms are “characterized by their X-ray powder diffraction patterns and/or by their solid state nuclear magnetic resonance spectra (NMR)” (A.R., Vol. 1, Tab 4, p. 91).
[17] Crystalline Form I atorvastatin is the subject of these proceedings. In the patent, the following is stated:
Form I atorvastatin consists of smaller particles and a more uniform size distribution than the previous amorphous product and exhibits more favorable filtration and drying characteristics. Additionally, Form I atorvastatin is purer and more stable than the amorphous product. (A.R., Vol. 1, Tab 4, p. 83)
[18] Form I is specifically claimed in Claims 1 to 9. Claims 1 to 5 describe the invention by setting out its reference peaks (referred to as 2θ values) on X-ray powder diffraction diffractograms. X-ray powder diffraction (XRPD) is a technique used to identify crystals and to determine crystal structure. When X-rays strike a sample of the material of interest, a pattern of peaks (so-called 2θ values) is produced which appear because the distances between atoms in a crystal are related to the wavelength of X-rays by Bragg’s law. The affidavit evidence of both Dr. Rodríguez-Hornedo and Dr. Myerson is to the effect that the XRPD data provides a reliable way of identifying different crystalline substances and distinguishing among them. As stated by Dr. Myerson, “XRPD produces a pattern of peaks that acts as a signature or fingerprint for that substance” (A.R., Vol. 4, Tab 9, p. 839).
[19] The only differences among the first five claims are the values of the XRPD peaks described. Claim 5 of the 018 Patent relates to Form I:
5. A crystalline Form I atorvastatin hydrate having an X-ray powder diffraction containing the following 2θ values measured using CuK% radiation: 9.150, 9.470, 10.266, 10.560, 11.853, 12.195, 17.075, 19.485, 21.626, 21.960, 22.748, 23.335, 23.734, 24.438, 28.915 and 29.234. (A.R., Vol. 1, Tab 4, p. 109)
[20] Claim 4 includes the same 2θ values as Claim 5, but rounded-off to one decimal place. Claims 1 to 3 of the 018 Patent are similar but include a subset of the 2θ values listed in Claim 5, rounded-off to one decimal place.
[21] The 018 Patent also claims Forms I, II and IV of atorvastatin calcium using another identification technique known as Carbon-13 Nuclear Magnetic Resonance (13C NMR). Rather than using X-rays as a means of characterizing a substance, 13C NMR uses magnetic fields. By exposing a substance to magnetic fields and applying specific techniques, the chemical environment surrounding each atom in the material can be determined. A 13C NMR produces a spectrum of chemical shifts. As stated by Dr. Rodríguez-Hornedo at p. 13 of her affidavit, “Different crystalline forms will have unique chemical shift spectra”. (A.R., Vol. 4, Tab 12, p. 1019)
[22] Form I is claimed in this manner in Claims 6 to 9. Typifying these 4 claims is Claim 8:
8. A crystalline Form I atorvastatin hydrate characterized by solid state 13C nuclear magnetic resonance having the following chemical shift differences between the lowest ppm resonance and other resonances: 3.9, 5.1, 18.9, 20.6, 26.1, 43.6, 46.8, 49.2, 51.8, 92.5, 96.9, 99.6, 102.2, 106.3, 108.2, 109.8, 113.6, 115.7, 138.0, 145.4, 157.1 and 161.5. (A.R., Vol. 1, Tab 4, pp. 109-110)
[23] Claims 6, 7 and 9 are similar. Claims 6 and 7 include a subset of the chemical shift differences listed in Claim 8. Claim 9 expresses the chemical shifts of crystalline Form I atorvastatin calcium in parts per million.
[24] There is no dispute that the invention embodied in Claims 1 to 9 is a hydrate. Each of Claims 1 to 9 refers solely to a “hydrate”. In contrast, the claims for each of Forms II and IV specifically include both a hydrate of the crystalline form and the anhydrous form. In addition to the clear words of Claims 1 to 9 (see, for example, the word “hydrate” in Claim 4), the patent discloses that:
Crystalline Form I atorvastatin contains about 1 to 8 mol of water. Preferably, Form I atorvastatin contains 3 mol of water. (A.R., Vol. 1, Tab 4, p. 100)
[25] Ranbaxy points to the different language in the two sets of Claims as being important. Claims 1 to 5 are directed to crystalline Form I atorvastatin hydrate having an XRPD “containing” specific 2θ peaks, whereas the words used in Claims 6 to 9 indicate that the patented Form I is “characterized by” certain 13C NMR shifts. In Ranbaxy’s view, this different language means that a person skilled in the art would understand that “crystalline Form I atorvastatin hydrate”, as that term is used in Claims 1 to 5, means more than the specified 2θ values. Rather, Ranbaxy submits, the Claims should be construed as referring to a certain polymorphic form of atorvastatin that, at its most basic level, has a unique and characteristic unit cell. I cannot agree with this construction.
[26] The main problem with Ranbaxy’s construction is that it incorrectly focuses on the word “containing” in each of Claims 1 to 5. Each of the Claims is to a material “having” an XRPD “containing” certain 2θ values. The phrase must be read in its entirety. Read as a whole, I have no doubt that the each of Claims 1 to 5 is defined by the relevant XRPD peaks. In other words, if the material is crystalline Form I atorvastatin hydrate with the specified XRPD peaks of Claims 1, 2, 3, 4 or 5, it is the substance claimed. A similar conclusion can be reached for Claims 6 to 9 with respect to 13C NMR spectroscopy.
[27] Even Ranbaxy appears to have accepted this conclusion when it states in its NOA that “Generally, the claims of the 018 Patent are directed to . . . Crystalline Form I of atorvastatin hydrate as defined by X-ray powder diffraction and solid state [13C] NMR spectroscopy” [Emphasis added.] (A.R., Vol. 2, Tab 6, p. 317)
[28] Relatively late, Ranbaxy has made the argument that it is an essential element of Claims 1 to 5 that the substance be a hydrate and that Pfizer must, in effect, prove separately that the form of crystalline Form I atorvastatin calcium used by Ranbaxy is a hydrate. While I agree (as does Pfizer) that Form I must be a hydrate, I do not see that this makes the hydrate an element to be separately determined. The invention is defined or controlled by either the XRPD peaks or the 13C NMR shifts. Those are the essential elements. This question is discussed later in these reasons.
[29] In conclusion, having reviewed the 018 Patent and considered the evidence before me, I believe that a person of ordinary skill in the art reading Claims 1 to 9 would understand that the material described as crystalline Form I atorvastatin calcium hydrate has one of the following:
o the specific set of 2θ values (or XRPD peaks) set out in Claims 1 to 5; or
o the specific 13C NMR shifts set out in Claims 6 to 9.
5.2 Is Ranbaxy’s NOA adequate?
[30] Pfizer notes that Ranbaxy does not raise the location of the manufacture and use of the intermediates as an alleged basis of non-infringement in its NOA. Thus, Pfizer submits that Ranbaxy’s NOA was insufficient to support this allegation.
[31] A party (second person) seeking an NOC from the Minister, must comply with ss. 5(1)(b)(iv) and 5(3)(a) of the NOC Regulations. In particular, under s. 5(3)(a), the second person must “provide a detailed statement of the legal and factual basis for the allegation”. There is copious jurisprudence that discusses what constitutes an acceptable or adequate NOA. In general terms, an NOA is adequate if it makes the patentee fully aware of the grounds on which the second person claims that the relevant patent will not be infringed if an NOC is issued by the Minister (see, for example, Pfizer Canada Inc. et al. v. Canada (Minister of Health), 2006 FCA 214 at para. 4, leave to appeal to S.C.C. refused, [2006] S.C.C.A. No. 335). The patentee must be able to decide whether or not to initiate a s. 6 proceeding under the NOC Regulations (AB Hassle v. Canada (Minister of National Health and Welfare) (2000), 7 C.P.R. (4th) 272 at para. 17 (F.C.A.)). In simple terms, whether an NOA is adequate will depend on whether the second person has provided the patentee with a sufficient understanding of the case it has to meet.
[32] Both parties referred to the following criteria set out in Pfizer Canada Inc. v. Apotex Inc., (2004), 31 C.P.R. (4th) 214 (F.C.T.D.), 2003 FC 1428 at para. 32, aff’d (2004), 38 C.P.R. (4th) 400 (F.C.A.):
In assessing the adequacy of the NOA, the following guidance can be taken from a number of decisions of the Federal Court of Appeal, including Bayer AG v. Canada (Minister of National Health and Welfare) (1993), 51 C.P.R. (3d) 329 (F.C.A.); Glaxo Group Ltd. v. Canada (Minister of National Health and Welfare) (2000), 6 C.P.R. (4th) 73 at 81 (F.C.T.D.), aff'd (2001) 11 C.P.R. (4th) 417 (F.C.A.);
· A bald assertion of non-infringement is insufficient.
· It is permissible for the second person to withhold certain information regarding its formulation until subsequent to a confidentiality order being in place.
· The NOA will be adequate if further disclosure elaborates on the basis for which the allegation of non-infringement was made such that there is sufficient evidence upon which to evaluate the allegation.
[33] The Ranbaxy NOA sets out Ranbaxy’s allegation of non-infringement in respect of the 018 Patent:
Specifically, Ranbaxy alleges that it does not make, construct, use or sell crystalline Form I atorvastatin hydrate, crystalline Form II atorvastatin or crystalline Form IV atorvastatin or their hydrates.
. . .
Ran-Atorvastatin does not contain crystalline Form I atorvastatin hydrate, nor is it made by a process in which crystalline Form I atorvastatin hydrate is used or produced. (A.R., Vol. 2, Tab 6, pp. 318-319 [Emphasis added.]
[34] There is no doubt that the allegations related to the 018 Patent could have been more informative – particularly with respect to the issue of whether the Ranbaxy intermediates could be shown to be hydrates. However, I believe that the words of the NOA are sufficient to allow Pfizer to know that the basis of the allegation was likely that Ranbaxy was not using a compound that could be identified as either crystalline Form I atorvastatin as defined in the 018 Patent or as a hydrate, as required by the 018 Patent.
[35] We must also look at the subsequent disclosures made after the Confidentiality Order was in place. On May 11, 2005, the manufacturing sites were disclosed to Pfizer. At that point, it should have been apparent to Pfizer that Ranbaxy was relying – at least in part – on the fact that all of the manufacturing steps were taking place in India.
[36] I am not persuaded that the NOA was inadequate. Even if I am wrong on this issue, because of my conclusion that the allegations with respect to the 018 Patent are not justified, the issue of adequacy of the NOA is not determinative.
5.3 Does the manufacture and use of the Ranbaxy intermediates infringe Claims 1 to 5 or Claims 6
to 9 of the 018 Patent?
[37] We know that Ranbaxy’s end-product is not crystalline Form I atorvastatin hydrate. Rather, the focus of this hearing is on the three Ranbaxy intermediates. The phrase “making, constructing, using or selling” in s. 5(1)(b)(iv) of the NOC Regulations has been held to be broad enough to include the use of patented substance at an intermediate stage, even if the intermediate ceases to exist once the final drug product is formed (see, for example, Abbott Laboratories v. Canada (Minister of Health) (2006), 350 N.R. 242 at paras. 15-17, 2006 FCA 187; Abbott Laboratories v. Canada (Minister of Health), 2007 FCA 73 at para. 4). Ranbaxy does not dispute this legal conclusion, provided that the intermediates are used in Canada.
[38] What then is the basis for Ranbaxy’s allegation that it does not infringe the 018 Patent? Having now reviewed all of the submissions and the arguments of Ranbaxy, I believe that the key arguments of Ranbaxy are the following:
The Ranbaxy intermediates do not fit the same unit cell as the patented Form I atorvastatin hydrate, meaning that Pfizer has not satisfied its burden to show that the Ranbaxy intermediates are the substances set out in Claims 1 to 9;
Pfizer has failed to show that Ranbaxy uses crystalline Form I atorvastatin hydrate; and
Ranbaxy manufactures or uses the intermediates in India, and not in Canada.
[39] While initially asserting that Claims 6 to 9 of the 018 Patent were invalid, Ranbaxy did not pursue this allegation in argument before me. Accordingly, for purposes of this Application, I will assume that all of the Claims in issue (Claims 1 to 9) of the 018 Patent are valid.
[40] On the issue of infringement, I begin with the testimony of Ranbaxy’s Senior Vice-President, Intellectual Property. During cross-examination, Mr. Deshmukh admitted that Ranbaxy uses crystalline Form I atorvastatin calcium to make the Ranbaxy product:
24 Q. One process [for making the Ranbaxy Product] uses Form I as an intermediate; is that correct?
A. The only one which is being currently commercialized uses Form I, as I understand it.
25 Q. Form I is the crystal form that is described in the Canadian Patent 2,220,018, correct?
A. To the best of my understanding, yes. (A.R., Vol. 12, Tab 17, pp. 3475-3476)
[41] Upon further questioning, it became clear that Mr. Deshmuck, apparently on behalf of Ranbaxy, his employer, took the position that the location of Ranbaxy’s operations is the basis of its allegation that it does not infringe the 018 Patent:
69 Q. To your knowledge, Form I is used in the process [for making the Ranbaxy Product], though, correct?
A. To my knowledge, what the intermediate is as I understand it is Form I.
70 Q. To your knowledge, then, the basis for this allegation [of non-infringement of the 018 Patent] is that the process is not run in Canada, correct?
A. Yes. [Emphasis added.]
(A.R., Vol. 12, Tab 17, pp. 3483-3484)
[42] This is, in effect, an admission that, but for the use of the intermediates in India, Ranbaxy would infringe the 018 Patent. In argument before me, Ranbaxy attempted to distance itself from the admissions by Mr. Deshmukh by arguing that he was not qualified to express this opinion. In spite of these arguments, I believe that it is compelling to hear from a senior executive of Ranbaxy that he acknowledges infringement but for the issue of the location of the Ran-Atorvastatin production facilities.
[43] In spite of this, I will now proceed to deal with the first two arguments by Ranbaxy on infringement.
5.3.1 Do the Ranbaxy intermediates consist of crystalline Form I atorvastatin calcium as defined by
the 018 Patent?
[44] The first question is whether the Ranbaxy intermediates (or any of them) fall within Claims 1 to 9 of the 018 Patent. The first aspect of this analysis is whether the sample data produced by Ranbaxy demonstrates a match with Pfizer’s crystalline Form I atorvastatin.
[45] As part of this proceeding, Ranbaxy produced XRPD data for four samples: one for each of the calcium seed and ATV-2P and two for ATV-2. Ranbaxy did not produce any 13C NMR data.
[46] At this point, the expert evidence becomes particularly helpful. I will begin with the evidence presented by the experts retained by Pfizer – Dr. Myerson and Dr. Rodríguez-Hornedo.
(a) Evidence of Drs. Myerson and Rodríguez-Hornedo
[47] Dr. Myerson used three analytical methods to identify the crystalline form of the Ranbaxy intermediates.
(i) Peak analysis – Dr. Myerson compared the 2θ values in each of the XRPD “peak search” tables produced by Ranbaxy to the 2θ values that are claimed in the 018 Patent;
(ii) Spectra overlay – Dr. Myerson visually compared Ranbaxy’s XRPD spectra relating to the intermediates isolated during the process of manufacturing Ran-Atorvastatin to the XRPD spectra included in the 018 Patent;
(iii) XRPD indexing – Using the data in the 018 Patent, Dr. Myerson calculated a unit cell for Form I atorvastatin calcium. Using the same calculation method, he analyzed the Ranbaxy XRPD data to determine whether the calculated unit cell was that of Form I atorvastatin calcium.
[48] Dr. Myerson was unequivocal that all three analyses lead to the same conclusion: the Ranbaxy intermediates are, or contain, crystalline Form I atorvastatin calcium. In his affidavit, Dr. Myerson summarized the results of his peak analysis in the following chart (A.R., Vol. 4, Tab 9, p. 851):
[Confidential pursuant to the Protective Order dated April 7, 2005.]
[49] Dr. Myerson’s visual overlay, although perhaps less scientific than the XRPD analysis, provides support for his peak analysis.
[50] With respect to the powder indexing he performed, Dr. Myerson concluded that:
the results for all four samples are almost identical to the results obtained for Form I atorvastatin, thus further confirming that these Ranbaxy samples are made up primarily [of] Form I. (A.R., Vol. 4, Tab 9, pp. 855-856)
His overall conclusion is that Ranbaxy’s Atorvastatin Seed, Atorvastatin Crude and Atorvastatin Calcium therefore fall within claims 1 to 5 of the 018 Patent. (A.R., Vol. 4, Tab 9, p. 856)
[51] Dr. Rodríguez-Hornedo conducted the same first two analyses as were carried out by Dr. Myerson and concluded as follows:
It is my view that the crystalline form of atorvastatin present in Ranbaxy’s Atorvastatin Calcium Seed, Atorvastatin Calcium Crude (ATV-2) and Atorvastatin Calcium (ATV-2P) is Form I, as claimed in claims 1-5 of the ‘018 patent. In my view, Ranbaxy’s atorvastatin seed and intermediates fall within claims 1 to 5 of the ‘018 patent. (A.R., Vol. 4, Tab 9, p. 1025)
(b) Evidence of Dr. Hollingsworth
[52] Dr. Hollingsworth, one of Ranbaxy’s experts, agreed, in his cross-examination, that the XRPD data from Ranbaxy’s samples have all of the 16 peaks claimed by the 018 Patent, within a typical margin of error of 0.2 degrees:
Q. Okay. So subject to your refusal to employ the typical 0.2 difference, you would agree that each of those four samples has all the 16 peaks of Form I in the 018 patent.
A.Yeah, there’s a match in the peak positions. That’s as far as I can – a match within 0.2 degrees, and that’s as far as I can go with that.
(A.R., Vol. 16, Tab 21, p. 5237)
[53] The “typical 0.2 difference” is a reference to the “typical” ± margin of error used in this type of analysis by persons skilled in the art and was used by Pfizer’s experts in their peak comparisons. In their view, the existence of differences of up to the 0.2 2θ is common and may be associated with instrument variability, sample preparation and preferred orientation (see, for example, p. 18, affidavit of Dr. Rodríguez-Hornedo (A.R., Vol. 4, Tab 12, p. 1024)). Dr. Hollingsworth is of the opinion that smaller errors than this “typical” value would be appropriate with respect to atorvastatin in order to uniquely identify it (A.R., Vol. 16, Tab 21, pp. 5235-5237).
[54] However, it is apparent, from reading Dr. Hollingsworth’s evidence in his affidavit and during cross-examination, that his disagreement was more fundamental than merely relying on a smaller error margin. Specifically, Dr. Hollingsworth does not believe that one should identify the Ranbaxy intermediates exclusively on the basis of a comparison of XRPD peaks (regardless of the margin of error). In his view, “the person skilled in the art would understand Form I atorvastatin to mean more than just its powder x-ray diffraction pattern and NMR peak positions, but more fundamentally its crystal structure, and in particular, its unit cell”. In his view, the crystalline Form I atorvastatin hydrate provided for in Claims 1 to 5 can only be identified by its “unique and characteristic unit cell”. As he states in his affidavit:
Solid-state forms of compounds that are crystalline have regularly repeating three-dimensional structures, the smallest unit of which is called the unit cell. The unit cell is the fundamental building block of any crystalline material, and it is characterized by the dimensions (in angstroms) along three axes . . . and the angles between these axes . . .
Crystal forms that have the same chemical composition are called polymorphs. In general, different polymorphs will have different unit cells, and the arrangement of atoms or molecules within those unit cells will be different. (A.R., Vol. 15, Tab 20, p. 4610)
[55] With this perspective, Dr. Hollingsworth, relying on data produced by Dr. Myerson in his XRPD indexing, carried out his own unit cell analysis or calculations. Specifically, Dr. Hollingsworth compared the unit cell that Dr. Myerson derived from the 018 Patent (Cell 1) to that derived from the Ranbaxy crude (ATV-2) sample (Cell 2). He concluded that there were significant differences between Cell 1 and Cell 2.
In light of the normal precision for unit cells determined from powder X-ray diffraction, it is my opinion that it would be clear to any person skilled in the art that these unit cell constants are significantly different and that it cannot be concluded on this basis that the 1118640 sample is the same as the polymorphic form described in the 018 Patent (A.R., Vol. 15, Tab 20, p. 4617) [emphasis in original].
[56] The difference in Cell 1 and Cell 2, he submits:
… calls in to question the validity of claiming a set of peak positions that contain little or no information about one of the three unit cell axes. This, in turn, calls into question the validity of simply comparing the peak positions in the Ranbaxy samples to those recited in the claims of the 018 and 455 patents. If the peaks from two dramatically different sets of cell constants give peaks that fall within the normally allowed error limits of the positions claimed in the two patents, then, as a matter of science, it is not possible to use those peak positions to distinguish these two materials. (A.R., Vol. 15, Tab 20, p. 4627)
[57] There are a number of problems with the evidence and arguments of Ranbaxy on this point.
[58] Although he qualified his responses, Dr. Hollingsworth acknowledged, during cross-examination, that XRPD is probably the most definitive method for identifying polymorphs and distinguishing among them (A.R., Vol. 16, Tab 21, p. 5259) and, further, that X-ray crystallographic methods which reflect differences in crystal structure can be definitive in such identification (A.R., Vol. 16, Tab 21, p. 5260). He also would not go so far as to opine that the Ranbaxy intermediate samples were not Form I. When asked directly, his response was, “It’s my opinion – that you can’t tell” (A.R., Vol. 16, Tab 21, p. 5232). This must be contrasted to the direct and unequivocal opinions of both Drs. Myerson and Rodríguez-Hornedo.
[59] Dr. Hollingsworth agreed that all XRPD data for the four samples provided by Ranbaxy (for the seed, ATV-2 and ATV-2P) showed a match in the 16 peak positions with the patented substance, within an accepted error of margin (A.R., Vol. 16, Tab 21, pp. 5236-5237, 5272-5275).
[60] When asked whether he was aware of any example where two samples shared 16 peaks and yet were different polymorphs, Dr. Hollingsworth’s response was only “I am trying to think of something, but nothing’s coming to mind right now” (A.R., Vol. 16, Tab 21, p. 5231). Indeed, the only example cited in his affidavit was an analysis (revealed through a literature search) of the two known forms of terephthalic acid where he noted “the inability to identify the two polymorphic forms of terephthalic acid based on a limited number of PXRD peaks” (A.R., Vol. 16, Tab 20, p. 4965). Apparently, in that example, although the three largest peaks matched in 20 values for both forms of the acid, the two forms were different. I have two difficulties in accepting this example as a reliable indicator that XRPD peak analysis can be misleading. The first is that the terephthalic acid comparison was carried out on only three peaks; in the case before me, the matching was consistent across 16 peaks. While I can accept that “different polymorphic forms of the same compound can, and often do, share many of the same peaks” (A.R., Vol. 16, Tab 20, p. 4965), this is a far cry from a situation where every one of the existing 16 peaks are almost identically aligned.
[61] Further, I note the limited nature of Dr. Hollingsworth’s analysis. Rather than conduct his own analysis on all four samples, he used the data produced by Dr. Myerson for one sample only: that being the ATV-2 or crude sample. As acknowledged by Dr. Hollingsworth, an impurity, as would likely exist in the ATV-2 sample, could add peaks to the XRPD (A.R., Vol. 16, Tab 20, p. 5284). In addition, Dr. Hollingsworth performed no other analysis that might have refuted or supported his finding. Specifically, he did not produce his own diffractograms or identify 13C NMR shifts. This limited approach raises significant doubt as to the reliance that should be placed on Dr. Hollingsworth’s conclusions.
(c) Ranbaxy intermediates and Claims 6 to 9
[62] I turn to consideration of Ranbaxy’s allegations related to Claims 6 to 9. In its NOA, Ranbaxy alleges that “it does not make, construct, use or sell crystalline Form I atorvastatin hydrate” (A.R., Vol. 2, Tab 6, p. 318). An essential element of Claims 6 to 9, as I determined above, is that the substance contain the specific 13C NMR shifts set out in those claims of the 018 Patent. As noted already, Ranbaxy did not produce any 13C NMR shift data for its intermediates. Pfizer argues that, because of Ranbaxy’s failure to produce these data, Ranbaxy has no basis to say that its intermediates do not infringe any of Claims 6 to 9.
[63] I agree to a point. Absent this evidence, Ranbaxy’s allegations of non-infringement of Claims 6 to 9 of the 018 Patent cannot be justified, unless Ranbaxy’s argument that Pfizer has not shown that Ranbaxy intermediates are hydrates can be sustained. The issue of whether the Ranbaxy intermediates are hydrates as contemplated by Claims 6 to 9 is discussed below.
(d) Conclusion on Ranbaxy intermediates
[64] In conclusion, I am not persuaded that Dr. Hollingsworth’s evidence should be accorded any significant weight.
[65] This, of course, leaves me with the expert evidence of Drs. Myerson and Rodríguez-Hornedo. As described above, each of these experts concluded that the Ranbaxy intermediates fall within Claims 1 to 5 of the 018 Patent. Their evidence supports a conclusion that, on a balance of probabilities, the Ranbaxy intermediates fall within Claims 1 to 5. In addition, the failure of Ranbaxy to provide 13C NMR data presents many problems in respect of Ranbaxy’s allegations. The issue that remains to be addressed is whether Pfizer has satisfied its burden of demonstrating that the Ranbaxy intermediates are crystalline Form I atorvastatin hydrates.
5.3.2 Does Ranbaxy use the hydrate as an intermediate in the making of Ran-Atorvastatin?
[66] Having established that the proper construction of Claims 1 to 9 limits the patent to the crystalline Form I atorvastatin calcium hydrate, the critical question is whether Pfizer has established on a balance of probabilities that Ranbaxy’s seed, ATV or ATV-2P are the crystalline Form I atorvastatin calcium hydrate.
[67] Ranbaxy argues that, even if it can be shown that it uses crystalline Form I atorvastatin calcium, Pfizer has failed to provide any evidence that Ranbaxy uses crystalline Form I atorvastatin hydrate. In summary form, Ranbaxy relies on the following evidence to support its argument:
Drs. Myerson and Rodríguez-Hornedo addressed only whether Ranbaxy makes crystalline Form I atorvastatin; their evidence made no observations on whether Ranbaxy makes or uses the hydrate. That is, neither witness considered the fact that it is an essential element of the Claims 1 to 9 that the material be a hydrate.
For a substance to be characterized as a hydrate, molecules of water must be trapped inside the crystal lattice; it is not enough that water merely be present. In support, Ranbaxy refers to the evidence of Dr. Myerson set out in his affidavit at paragraph 28 and that of Dr. Rodríguez-Hornedo where she confirmed in cross-examination that water molecules would be a part of the lattice structure of a hydrate (A.R., Vol. 4, Tab 13, p. 1100) and stated, in her affidavit:
When a crystal is formed, sometimes molecules of the solvent are incorporated into the crystal lattice. When this happens, the crystal form is called a “solvate” (or a “hydrate” when the solvent is water). If no molecules of the solvent are present inside the crystal lattice, the crystalline form is called “anhydrous”, although water can sometimes be associated with anhydrous crystalline forms by absorption of water molecules on to the surface of the crystal. (A.R., Vol. 4, Tab 12, p. 1017)
The best evidence (in Ranbaxy’s vie