Millennium Pharmaceuticals, Inc and another v Zyfas Medical Co (sued as a firm) [2023] SGHC 360

Case Details

• Citation: [2023] SGHC 360
• Court: General Division of the High Court of the Republic of Singapore
• Decision Date: 29 December 2023
• Coram: Dedar Singh Gill J
• Case Number: Suit No 817 of 2019
• Hearing Date(s): 19–20, 22, 26–27 October 2021, 27 February, 5 May 2023
• Claimants / Plaintiffs: (1) Millennium Pharmaceuticals, Inc; (2) Johnson & Johnson Pte Ltd
• Respondent / Defendant: Zyfas Medical Co (sued as a firm)
• Counsel for Claimants: Suhaimi Bin Lazim, S Siddharth Sriram, Shahera Safrin (Mirandah Law LLP)
• Counsel for Respondent: Wong Siew Hong, Poonaam Bai (Eldan Law LLP)
• Practice Areas: Intellectual Property — Patents and inventions — Validity; Patents — Validity and Infringement

Summary

The decision in Millennium Pharmaceuticals, Inc and another v Zyfas Medical Co (sued as a firm) [2023] SGHC 360 represents a significant addition to Singapore’s patent jurisprudence, particularly concerning the validity and infringement of pharmaceutical process patents. The dispute centered on two patents held by the first plaintiff, Millennium Pharmaceuticals, Inc, which disclosed methods for the large-scale synthesis of boronic acid and ester compounds essential for the production of bortezomib. Bortezomib is a potent proteasome inhibitor used primarily in the treatment of multiple myeloma and mantle cell lymphoma. The second plaintiff, Johnson & Johnson Pte Ltd, as the exclusive licensee, distributed the brand-name drug Velcade in Singapore. The defendant, Zyfas Medical Co, sought to enter the market with a generic version, "Myborte," manufactured by Dr. Reddy’s Laboratories (“DRL”).

The litigation involved a dual-track challenge: the plaintiffs alleged infringement of their patents under the Patents Act, while the defendant counterclaimed for the revocation of the patents on grounds of lack of novelty, lack of inventive step, and insufficiency. The two patents in suit were Singapore Publication No. SG 151322 (“SG 322”) and Singapore Application No. SG 10201600029P (“SG 29P”). SG 322 focused on improvements to the Matteson Homologation process through the use of specific coordinating ether solvents with low water miscibility. SG 29P claimed a "convergent synthesis" approach for the large-scale manufacture of bortezomib, which the plaintiffs argued offered superior purity and process control compared to traditional linear synthesis methods.

The High Court, presided over by Dedar Singh Gill J, navigated the extreme technical complexity of the case with the assistance of an independent court assessor, Professor Paul Sharratt. The court’s analysis of SG 322 resulted in a finding of validity, as the specific solvent parameters were deemed non-obvious to a person skilled in the art at the priority date. However, the infringement claim for SG 322 failed because the plaintiffs could not prove that DRL’s manufacturing process actually utilized the claimed solvent conditions. Regarding SG 29P, the court held the patent invalid for lack of inventive step, concluding that the shift from linear to convergent synthesis was a matter of routine optimization for a skilled chemist tasked with scaling up production.

This judgment underscores the rigorous evidentiary burden placed on patent holders in process infringement suits, especially when the manufacturing occurs abroad and the "reversal of burden of proof" under s 68 of the Patents Act is contested. It also clarifies the "routine optimization" threshold for inventive steps in chemical synthesis. By dismissing the infringement claims and revoking one of the two patents, the court balanced the protection of pharmaceutical innovation with the statutory pathways for generic competition, providing a detailed roadmap for practitioners handling complex IP disputes in the life sciences sector.

Timeline of Events

1. 25 June 1985: Priority date or related historical reference in prior art (Matteson 1985).
2. 27 August 1985: Related historical reference in prior art.
3. 4 July 1989: Related historical reference in prior art.
4. 9 May 1996: Related historical reference in prior art.
5. 14 July 1998: Related historical reference in prior art.
6. 28 February 2001: Related historical reference in prior art.
7. 12 October 2001: Related historical reference in prior art.
8. 24 April 2003: Related historical reference in prior art.
9. 15 March 2005: Related priority date reference.
10. 24 March 2005: Filing date for SG 322 (Singapore Publication No. SG 151322).
11. 31 July 2012: Grant date for SG 322.
12. 19 February 2018: Related procedural date in the lead-up to litigation.
13. 5 July 2019: Related procedural date.
14. 11 July 2019: Related procedural date.
15. 15 July 2019: Related procedural date.
16. 24 July 2019: Related procedural date.
17. 30 July 2019: Related procedural date.
18. 31 July 2019: Related procedural date.
19. 2 August 2019: Related procedural date.
20. 19 August 2019: The first plaintiff commenced Suit No 817 of 2019 in the High Court.
21. 27 August 2019: Related procedural date.
22. 23 October 2019: Related procedural date.
23. 12 November 2019: Related procedural date.
24. 21 November 2019: Related procedural date.
25. 29 November 2019: Related procedural date.
26. 29 January 2020: Related procedural date.
27. 16 March 2020: Related procedural date.
28. 2 April 2020: Related procedural date.
29. 8 April 2020: Related procedural date.
30. 9 April 2020: Related procedural date.
31. 29 May 2020: Related procedural date.
32. 9 June 2020: Related procedural date.
33. 12 June 2020: Related procedural date.
34. 22 June 2020: Related procedural date.
35. 15 July 2020: Related procedural date.
36. 21 July 2020: Related procedural date.
37. 26 August 2020: Related procedural date.
38. 27 August 2020: Related procedural date.
39. 21 December 2020: Related procedural date.
40. 19 March 2021: Related procedural date.
41. 10 May 2021: Related procedural date.
42. 11 May 2021: Related procedural date.
43. 12 May 2021: Related procedural date.
44. 30 May 2021: Related procedural date.
45. 8 June 2021: Related procedural date.
46. 19 June 2021: Related procedural date.
47. 6 July 2021: Related procedural date.
48. 7 July 2021: Related procedural date.
49. 13 August 2021: Related procedural date.
50. 15 August 2021: Related procedural date.
51. 16 August 2021: Related procedural date.
52. 27 August 2021: Related procedural date.
53. 9 September 2021: Related procedural date.
54. 13 September 2021: Related procedural date.
55. 14 September 2021: Related procedural date.
56. 7 October 2021: Related procedural date.
57. 11 October 2021: Related procedural date.
58. 19–20, 22, 26–27 October 2021: Substantive hearing dates for the liability phase.
59. 27 February, 5 May 2023: Further hearing dates.
60. 29 December 2023: Final judgment delivered by Dedar Singh Gill J.

What Were the Facts of This Case?

The dispute in this case arose within the highly regulated and technically complex pharmaceutical industry, specifically concerning the synthesis of bortezomib. Bortezomib is a proteasome inhibitor used for treating multiple myeloma and mantle cell lymphoma. The first plaintiff, Millennium Pharmaceuticals, Inc ("Millennium"), is the registered proprietor of two Singapore patents: SG 151322 (“SG 322”) and SG 10201600029P (“SG 29P”). The second plaintiff, Johnson & Johnson Pte Ltd (“J&J”), is the exclusive licensee of these patents in Singapore and the distributor of the brand-name drug Velcade.

The defendant, Zyfas Medical Co (“Zyfas”), is a Singapore-registered firm involved in the wholesale and distribution of pharmaceutical products. Zyfas obtained registration under the Health Products Act (Cap 122D) for a therapeutic product named "Myborte Powder for Solution for Injection 3.5mg/vial" (the “Alleged Infringing Product”). Myborte is a generic version of bortezomib manufactured by Dr. Reddy’s Laboratories (“DRL”), an Indian pharmaceutical company. The plaintiffs alleged that the manufacture of Myborte by DRL utilized processes that infringed the asserted claims of SG 322 and SG 29P.

The technical core of the case involved the large-scale synthesis of boronic acid and ester compounds, which are critical intermediates in the production of bortezomib. SG 322, filed on 24 March 2005 and granted on 31 July 2012, specifically addresses improvements to the Matteson Homologation process. This process is used to create chiral boronic esters. The patent claims that by using a "coordinating ether solvent" with "low miscibility in water" (defined as less than about 10% by weight), the efficiency and purity of the synthesis are significantly improved, particularly for large-scale industrial applications. The plaintiffs argued that this specific solvent choice was a non-obvious improvement that allowed for better separation and higher yields, with technical evidence citing purities as high as 98.0% and 99.5% and water content as low as 0.1% or 0.5% in the reaction mixtures.

SG 29P, on the other hand, focuses on a "convergent synthesis" approach. In organic chemistry, a convergent synthesis involves preparing different parts of a complex molecule separately and then joining them together in a final step. This is contrasted with "linear synthesis," where the molecule is built step-by-step from a single starting point. The plaintiffs contended that this convergent method was superior for the large-scale manufacture of bortezomib, offering advantages in terms of purity and process control that were not present in the prior art linear methods.

The defendant denied infringement and counterclaimed for the revocation of both patents. Zyfas asserted that the patents were invalid for lack of novelty and lack of inventive step, relying on various prior art documents, including the "Matteson and Majumdar" paper from the Journal of the American Chemical Society (1980) Vol 102. The defendant argued that the methods claimed were either already known or were obvious variations of existing chemical techniques that a person skilled in the art would have arrived at through routine optimization.

Given the technical complexity, the parties agreed to the appointment of a court assessor, Professor Paul Sharratt, under s 10A(1) of the Supreme Court Judicature Act (Cap 322) and O 33 r 4 of the ROC. Professor Sharratt, a specialist in chemical and pharmaceutical process development, assisted the court in reviewing the technical evidence. The litigation proceeded on a bifurcated basis, with the present judgment focusing on liability. The procedural history was marked by numerous interlocutory summonses (including Summons No 430/2020, 437/2020, 1716/2020, 2121/2020, 2368/2020, and 2850/2020) and a prior related decision in [2020] SGHC 28, which dealt with the defendant's omission to disclose certain matters during the registration process.

What Were the Key Legal Issues?

The resolution of this dispute required the court to address several fundamental issues of patent law, framed by the statutory requirements of the Patents Act. The primary legal issues were as follows:

• Validity of SG 322 and SG 29P: The court had to determine whether the patents met the requirements for patentability under s 13(1) of the Patents Act. This involved:
  • Novelty (s 14): Whether the inventions formed part of the "state of the art" before the priority date.
  • Inventive Step (s 15): Whether the inventions were obvious to a person skilled in the art, applying the Windsurfing/Pozzoli test.
  • Sufficiency of Disclosure (s 25(4) and s 80(1)(c)): Whether the patent specifications disclosed the invention in a manner clear and complete enough for it to be performed by a person skilled in the art.
• Infringement (s 66): Whether the defendant’s supply of Myborte infringed the asserted claims of SG 322 and SG 29P. This required a comparison between the patented processes and the manufacturing process used by DRL.
• Reversal of Burden of Proof (s 68): Whether the subject matter of the patents was a "new product," thereby shifting the burden to the defendant to prove that the Alleged Infringing Product was not made by the patented process.
• Standing of the Second Plaintiff (s 74): Whether J&J qualified as an "exclusive licensee" under s 2(1) of the Patents Act, granting it the right to bring proceedings for infringement.
• Claim Construction: The proper interpretation of key terms in the patent claims, such as "coordinating ether solvent" and "low miscibility in water" in SG 322, using a purposive approach.
• The Role of the Court Assessor: The extent to which the court should rely on the technical findings of Professor Paul Sharratt while maintaining its role as the ultimate arbiter of legal issues.

Each of these issues was critical. The validity challenge struck at the heart of the plaintiffs' monopoly, while the infringement analysis tested the practical reach of process patents against foreign-manufactured generics. The application of s 68 was particularly significant, as it could determine the outcome where direct evidence of a foreign manufacturing process is difficult to obtain.

How Did the Court Analyse the Issues?

The court’s analysis was exhaustive, spanning over 200 pages and integrating complex chemical evidence with established legal doctrines. The analysis began with the foundational task of claim construction and the identification of the "person skilled in the art."

Claim Construction and the Skilled Addressee

Following the principles in Currency Choice Pte Ltd v Main-Line Corporate Holdings Ltd and another appeal [2008] 1 SLR(R) 335, the court adopted a purposive construction of the patent claims. The objective was to determine what the person skilled in the art would have understood the patentee to be using the language of the claim to mean. The court defined the "skilled addressee" for both patents as an individual with a graduate degree in chemistry (either a Masters of Science or a PhD) and at least five years of experience in organic synthesis in an academic or industrial laboratory. This profile was essential for evaluating both the inventive step and the sufficiency of the disclosure.

Validity of SG 322: The Windsurfing/Pozzoli Test

The court applied the four-step Windsurfing test (as refined in Pozzoli SPA v BDMO SA [2007] RPC 729) to assess the inventive step of SG 322:

1. Identify the notional "person skilled in the art" and the relevant common general knowledge.
2. Identify the inventive concept of the claim.
3. Identify differences between the "state of the art" and the inventive concept.
4. Determine if those differences constitute steps which would have been obvious to the person skilled in the art.

The inventive concept of SG 322 was the use of a coordinating ether solvent with low miscibility in water (less than about 10% by weight) to improve the Matteson Homologation process for large-scale synthesis. The court found that while the Matteson Homologation was known, the specific selection of solvents with these miscibility parameters to solve the problems of large-scale production (such as the need for efficient aqueous washes and high purity) was not obvious. The defendant failed to show that the prior art, such as the Matteson and Majumdar (1980) paper, taught or suggested this specific combination. Consequently, SG 322 was held to be valid.

Infringement of SG 322 and Section 68

The infringement analysis for SG 322 turned on whether the plaintiffs could prove that DRL used the patented process. The plaintiffs sought to invoke s 68(1) of the Patents Act, which reverses the burden of proof if the product obtained by the patented process is "new." The court referred to Merck & Co Inc v Pharmaforte Singapore Pte Ltd [1999] 3 SLR(R) 1072, noting that "new product" includes anything not known in the state of the art. However, the court found that the plaintiffs failed to provide sufficient evidence to show that DRL’s process actually utilized the specific solvent conditions claimed. Even if the burden had shifted, the technical evidence regarding DRL's process did not align with the narrow construction of the "coordinating ether solvent" and "low miscibility" requirements. The court noted:

"Although SG 322 is valid, in my view, the plaintiffs have not shown how any of the asserted claims have been infringed by DRL’s manufacturing process as presented." (at [289])

Validity of SG 29P: Routine Optimization

For SG 29P, the court’s focus was on the inventive step of "convergent synthesis." The court compared this to the prior art linear synthesis methods. Drawing on Union Carbide Corp v BP Chemicals [1998] RPC 1 and Dyson Appliances Ltd v Hoover Ltd [2001] RPC 26, the court analyzed whether the shift to a convergent approach was an "inventive" leap or merely "routine optimization."

The court, assisted by Professor Sharratt, concluded that a skilled chemist tasked with scaling up the production of a complex molecule like bortezomib would naturally consider a convergent approach to improve efficiency, yield, and purity. This was considered a standard strategy in organic chemistry. The specific reagents and steps used in SG 29P were found to be either known or obvious variations of existing techniques. The court held that the differences between the prior art and SG 29P did not require any degree of invention. Thus, SG 29P was declared invalid for lack of inventive step under s 15.

The Role of the Court Assessor

The court emphasized the utility of the assessor under s 10A(1) of the Supreme Court Judicature Act. Professor Sharratt’s role was to explain the technical evidence and provide an independent perspective on what a skilled addressee would have known. The court adopted the view from Genetic Institute Inc v Kirin-Amgen Inc (No 2) (1997) 149 ALR 247 that the assessor does not give judgment but assists the judge in understanding the technical matrix. The court was careful to ensure that the ultimate legal conclusions on validity and infringement remained the province of the judge.

What Was the Outcome?

The High Court reached a split decision regarding the two patents, ultimately resulting in the dismissal of the plaintiffs' infringement claims and the partial success of the defendant's counterclaim for revocation. The court's orders were as follows:

• SG 322: The court found that SG 322 was valid. It met the requirements for novelty and inventive step, and the disclosure was sufficient. However, the court found that the plaintiffs failed to prove infringement. The evidence did not establish that the manufacturing process used by DRL for the Alleged Infringing Product fell within the scope of the asserted claims of SG 322.
• SG 29P: The court found that SG 29P was invalid for lack of inventive step under s 15 of the Patents Act. The move from linear to convergent synthesis was deemed routine optimization for a person skilled in the art. Consequently, the defendant's counterclaim for the revocation of SG 29P was successful.
• Costs: The court awarded costs to the defendant, as the overall result was in its favor (the infringement claims were dismissed and one patent was revoked).

The operative conclusion of the judgment is captured in paragraph 289:

"For the reasons above, I find that SG 29P is invalid for lack of inventive step. Although SG 322 is valid, in my view, the plaintiffs have not shown how any of the asserted claims have been infringed by DRL’s manufacturing process as presented. I award costs to the defendant. Costs are to be agreed, or otherwise taxed." (at [289])

The court's decision effectively cleared the path for the defendant to continue the distribution of Myborte in Singapore, provided it does not infringe the valid SG 322 patent through a different process. The plaintiffs' attempt to maintain a monopoly over the convergent synthesis of bortezomib was rejected, reinforcing the principle that standard chemical strategies cannot be monopolized unless they involve a truly inventive step beyond routine optimization.

Why Does This Case Matter?

The judgment in Millennium Pharmaceuticals v Zyfas Medical Co is a landmark decision for several reasons, impacting both the doctrinal landscape of Singapore patent law and the practical strategies of IP practitioners.

Refining the "Inventive Step" in Chemical Synthesis

The case provides a clear application of the Windsurfing/Pozzoli test to complex chemical processes. Specifically, it clarifies the boundary between "invention" and "routine optimization." By holding that the shift from linear to convergent synthesis was obvious, the court has set a high bar for process patents that rely on standard chemical strategies. This prevents the "evergreening" of patents through minor, predictable improvements in manufacturing efficiency. Practitioners must now ensure that any claimed process improvement offers a technical advantage that is not just desirable but also non-obvious to a skilled chemist.

The Evidentiary Burden in Process Infringement

The failure of the infringement claim for SG 322 highlights the significant evidentiary hurdles faced by patent proprietors. Even with the potential for a reversal of the burden of proof under s 68, the court required precise alignment between the patented claims and the defendant's actual process. This case serves as a warning that general allegations of infringement are insufficient; proprietors must be prepared to provide detailed technical evidence, often involving discovery of foreign manufacturing processes, to succeed.

The Use of Court Assessors

This is one of the most prominent recent examples of the use of a court assessor under s 10A(1) of the Supreme Court Judicature Act. The judgment demonstrates how an assessor can help the court navigate "dense" technical evidence without usurping the judicial function. For practitioners, this highlights the importance of framing technical arguments in a way that is both scientifically sound (to satisfy the assessor) and legally relevant (to satisfy the judge).

Standing of Exclusive Licensees

The court's consideration of J&J's standing under s 74 and s 2(1) of the Patents Act reaffirms the rights of exclusive licensees to participate in infringement proceedings. This is crucial for the pharmaceutical industry, where the patent owner and the local distributor are often different entities. It ensures that the party with the most direct commercial interest in the Singapore market has the standing to protect its exclusivity.

Impact on Generic Entry

Finally, the case illustrates the balance the Singapore courts strike between protecting innovation and facilitating generic competition. By revoking an invalid patent and strictly construing the valid one, the court ensured that generic manufacturers are not unfairly blocked by "weak" patents, while still upholding the validity of genuine technical improvements. This provides greater certainty for generic pharmaceutical companies looking to enter the Singapore market.

Practice Pointers

• Drafting Specificity: When drafting process claims, practitioners should include specific parameters (e.g., solvent miscibility, purity levels) that clearly distinguish the invention from routine optimization. Vague claims are more susceptible to lack of inventive step challenges.
• Evidence of Infringement: In process patent cases, especially those involving foreign manufacturers, practitioners should prioritize obtaining direct evidence of the manufacturing process early in the litigation. Relying solely on s 68's reversal of burden is risky if the "newness" of the product or the process's uniqueness is contested.
• Engaging with Assessors: When a court assessor is appointed, counsel should ensure that their expert witnesses are prepared to address the assessor's technical queries directly. The assessor's view of the "skilled addressee" and "common general knowledge" will be highly influential.
• Bifurcation Strategy: The use of bifurcation (liability vs. quantum) in this case allowed the court to focus on the complex technical issues of validity and infringement without being distracted by damages calculations. This is often an efficient approach in IP litigation.
• Routine Optimization Defense: For defendants, the "routine optimization" argument is a powerful tool against process patents. Demonstrating that a change (like moving to convergent synthesis) is a standard industry practice can lead to the revocation of the patent.
• Purposive Construction: Always frame claim construction arguments within the "purposive" framework of First Currency. The focus should be on what the skilled addressee would understand, not just a literal reading of the text.
• Section 68 Nuances: Be aware that s 68 only applies if the product is "new." If the product (e.g., bortezomib) is already known in the state of the art, the burden of proof for infringement remains firmly with the plaintiff.

Subsequent Treatment

As a decision delivered in late 2023, [2023] SGHC 360 remains a primary authority on the application of the Windsurfing/Pozzoli test to pharmaceutical process patents and the role of court assessors in Singapore. It has not yet been significantly distinguished or overruled by the Court of Appeal, and it stands as a robust example of the High Court's approach to balancing patent monopolies with generic competition in the life sciences sector.

Legislation Referenced

• Health Products Act (Cap 122D)
• Patents Act (Cap 221) [specifically ss 2(1), 13(1), 14, 15, 17, 25(4), 66(1), 68(1), 74, 75, 80(1)]
• Supreme Court Judicature Act (Cap 322) [specifically s 10A(1)]
• Rules of Court (ROC) [specifically O 33 r 4]

Cases Cited

• Applied / Followed:
  • Windsurfing International Inc v Tabur Marine (Great Britain) Ltd [1985] RPC 59
  • Pozzoli SPA v BDMO SA [2007] RPC 729
  • Currency Choice Pte Ltd v Main-Line Corporate Holdings Ltd and another appeal [2008] 1 SLR(R) 335
  • IIa Technologies Pte Ltd v Element Six Technologies Ltd [2023] 1 SLR 987
  • Sunseap Group Pte Ltd and others v Sun Electric Pte Ltd [2019] 1 SLR 645
• Considered / Referred to:
  • [2020] SGHC 28
  • Zyfas Medical Co (sued as a firm) v Millennium Pharmaceuticals, Inc [2020] 2 SLR 1044
  • Lee Tat Cheng v Maka GPS Technologies Pte Ltd [2018] 1 SLR 856
  • FE Global Electronics Pte Ltd v Trek Technology (Singapore) Pte Ltd [2006] 1 SLR(R) 874
  • Ng Kok Cheng v Chua Say Tiong [2001] 2 SLR(R) 326
  • Rohm and Haas Electronic Materials CMP Holdings, Inc v NexPlanar Corp [2018] 5 SLR 180
  • Merck & Co Inc v Pharmaforte Singapore Pte Ltd [2000] 2 SLR(R) 708
  • ASM Technology Singapore Pte Ltd v Towa Corp [2018] 1 SLR 211
  • Mühlbauer AG v Manufacturing Integration Technology Ltd [2010] 2 SLR 724
  • ASM Assembly Automation Ltd v Aurigin Technology Pte Ltd [2010] 1 SLR 1
  • Union Carbide Corp v BP Chemicals [1998] RPC 1
  • Dyson Appliances Ltd v Hoover Ltd [2001] RPC 26
  • Conor Medsystems Inc v Angiotech Pharmaceuticals Inc [2008] UKHL 49
  • Actavis Group PTC EHF v ICOS Corporation [2019] UKSC 15
  • Genetic Institute Inc v Kirin-Amgen Inc (No 2) (1997) 149 ALR 247

Source Documents

• Original judgment PDF: Download (PDF, hosted on Legal Wires CDN)
• Official eLitigation record: View on elitigation.sg