Summary: The judgement offers a nuanced interpretation of patent law, particularly the concepts of obviousness under Section 64(1)(f) of the Patents Act, 1970..
In a significant ruling, the Delhi High Court’s Division Bench, comprising Justice C. Hari Shankar and Justice Ajay Digpaul, recently dismissed F. Hoffmann-La Roche AG’s appeal against Natco Pharma Ltd, involving the alleged infringement of Roche’s Indian patent IN 334397, which claims Risdiplam, an oral drug used to treat Spinal Muscular Atrophy (SMA).
Background
Indian Patent No. IN 334397, titled “Compounds for Treating Spinal Muscular Atrophy”, was granted to Roche in 2020 and is valid until May 2035. The drug in question is used to treat Spinal Muscular Atrophy (SMA), a rare genetic neuromuscular disorder. Roche markets Risdiplam globally under the brand name Evrysdi.
Roche filed a patent infringement suit after discovering Natco’s plans to manufacture and sell a generic version of Risdiplam. Natco had listed Risdiplam under its “APIs under development” and filed a separate patent application in 2022 for improved process to manufacture the drug. Roche sought an interim injunction to restrain Natco from launching its version, alleging patent infringement.
Natco challenged the validity of Roche’s patent under Section 107, read with Section 64(1)(e) (lack of novelty) and Section 64(1)(f) (obviousness) of the Patents Act, 1970. The case proceeded before the Delhi High Court, first before a single judge and later in appeal before the division bench.
Findings of the Single Judge Bench
In its March 2025 order[1], the single judge bench of the Delhi High Court declined to grant Roche an interim injunction after closely examining Natco’s challenge to the validity of ‘Roche’s’ Suit Patent. It found credible grounds for invalidity, on grounds of anticipation and obviousness. On anticipation, the Court noted that both the genus patent (WO”’916) and the Suit Patent claimed the same compound and has four common inventors. Further, the learned Single Judge, in the impugned judgment, accepted the respondent’s contention that Risdiplam, which was claimed in the Suit Patent, was covered in the broad Markush formulae of WO’916. Additionally, the learned Single Judge found Claim 1 in the suit patent, which claimed Risdiplam, to be obvious vis-à-vis prior art in the form of Compound 809 of WO’916 (which corresponds to US Patent 958695515). The molecular formulae of Risdiplam and Compound 809 in WO’916/ US 955 have been set out side by side as follows:
Given the life-threatening nature of SMA, limited treatment options, and public interest, the single Judge concluded that granting an injunction was justified in the interest of affordable access.
Findings of the Division Bench
The order of the Single Judge was appealed. The findings of the appeal decision are discussed below:
Standard of appellate review:
The Court reiterated that under Wander[2] (and as recently reiterated by the Supreme Court), an appellate court should not substitute its discretion for that of the trial judge unless the exercise of discretion was arbitrary, capricious, or legally incorrect. Accordingly, the High Court confined itself to whether the Single Judge applied correct legal principles rather than re‑assessing the factual record afresh.
Statutory framework:
The Court set out Sections 48, 64 and 107 of the Patents Act. Section 48 confers exclusive rights on a patentee. Further, Section 64(1)(e) and (f) set out revocation grounds for lack of novelty (prior disclosure) and obviousness (lack of inventive step). Lastly, Section 107 makes available revocation grounds as defences in infringement suits.
Section 64(1)(e) (novelty / disclosure):
The DB expressed reservations in agreeing with the single judge’s findings on anticipation and treating coverage alone as sufficient for anticipation. It carefully distinguished “coverage” by a Markush genus from “disclosure”, and noted that the Single Judge found Risdiplam ‘disclosed’ in WO’916/US’955 because the appellant had alleged infringement of both genus and species patents in the US and also because of various filings and statements, made by the Patentee Roche in the Patent Term Extension (PTE) application filed in the US (which allows patent holders to recover time lost during regulatory approval for drugs and medical devices) and its listing in the FDA’s Orange Book (which catalogues patents covering an approved drug product), among others. The Division Bench observed that there is judicial divergence on whether mere coverage equals disclosure. The Court expressed reservations about treating coverage alone as sufficient for disclosure and said it was not prima facie inclined to hold that Section 64(1)(e) had been made out merely basis the genus documents WO’916/US‘955. However, it accepted that the Single Judge’s findings were in sync with the law laid down in Astrazeneca[3]. The DB, however, did not address whether Astrazeneca was correctly decided, and held that it should be examined during a more appropriate case.
The Bench further noted that any manufacture or sale by a defendant of a compound falling within the broad coverage of the Markush formulation in the genus patent would entitle the holder of the genus patent to sue for infringement. Accordingly, if the compound falls within the broad coverage of Markush formulation in the genus patent and is also specifically claimed in a species patent, a valid infringement claim may lie against a defendant who markets and sells the said product without a licence from the holder of both the genus and the species patents.
The Bench further observed that Supreme Court’s dismissal of the associated SLP did not constitute an affirmation on merits. The Court left open the question for future consideration.
Section 64(1)(f) (obviousness):
The Court concurred with the Single Judge’s prima facie finding that Risdiplam is obvious over Compound 809 in WO’916/US’955. The division bench observed that the only structural distinction between Compound 809 and Risdiplam is the substitution of a -CH radical by a Nitrogen (-N) atom at the circled junction, otherwise the structures are identical (as set out in the impugned judgment), as can be observed below:
Further, it noted that WO’916 discloses many exemplified compounds (835) in which nitrogen appears in various ring positions. Such disclosures clearly indicate that nitrogen substitution in the relevant positions was contemplated in the genus patent. Further, Imidazo [1,2‑b] pyridazinyl is specifically disclosed in WO’916 for the R2 radical, which supports a pathway from Compound 809 to Risdiplam. The Bench observed that disclosures in WO’916 provided sufficient guidance for persons skilled in the art — the inventors of WO’916 — to make the necessary substitution of -CH with -N, in the exemplified Compound 809, and thus obtain Risdiplam.
Further, DB also noted that Grimm’s Hyride Displacement Law (table of isosteres) places N and CH in the same vertical column and treats them as bioisosteres. Hence, substitution of CH with N would be a recognised medicinal‑chemistry modification, and bioisosteric replacement, and it would be obvious to a person skilled in medicinal chemistry. The table is as follows:
The court also observed that four lead inventors of the genus and species patents were common. It also accepted the “person in the know” idea (as articulated in Astrazeneca[4]) that where inventors overlap, an inventor’s perspective makes the pathway to the species easier to discern and increases the plausibility of obviousness. Basis this, the Court found the Single Judge’s conclusion that the respondent had made out a credible challenge to the suit patent under Section 64(1)(f) to be legally sound and not susceptible to interference within the Wander constraints.
Interaction of the two grounds:
The Court explained that it was not necessary to fully resolve the Section 64(1)(e) question because a credible challenge under any clause of Section 64 (including 64(1)(f)) suffices as a defence at this prima-facie stage. As the Court agreed with the Single Judge on the ground of ‘obviousness’ under Section 64(1)(f), its injunction application dismissal could be upheld on that ground alone.
On the precedential effect of Astrazeneca[5] and other filings:
The Court noted that Astrazeneca[6] influenced the Single Judge because the appellants had sued in the US alleging infringement of both genus and species patents. The Division Bench expressed doubts about Astrazeneca’s[7] broader legal proposition. The Bench observed that there were doubts on whether the principle laid down in Astrazeneca[8] was correct in law. It noted that infringement is predicated on coverage, whereas invalidity anticipation is predicated on disclosure, a distinction which, in its view, the Astrazeneca[9] decision overlooks.
Application of interlocutory standards:
The Court reiterated that in the interlocutory stage, the defendant needs to make out only a credible (not cast‑iron) challenge to validity. The Single Judge had found such a credible challenge on grounds of obviousness.
Implications
The judgment has notable implications for pharmaceutical patent litigation in India. It focuses on obviousness, indicating that minor structural modifications to known compounds may not meet the inventive step threshold. Hence, Natco was permitted to launch Risdiplam’s generic version at ₹15,900 per bottle, significantly lower than Roche’s listed price of ₹6 lakh. While some view this as a move toward improving patient access to affordable medication, others emphasise the importance of strong intellectual property protection in encouraging pharmaceutical innovation, attracting investment in drug development, and supporting broader healthcare goals under initiatives like Viksit Bharat.
[1] 2025 DHC 1907
[2] Wander Ltd v Antox India (1990 Supp SCC 727)
[3] Astrazeneca AB v Intas Pharmaceuticals Ltd (2020) 84 PTC 326 (Del)
[4] supra
[5] supra
[6] supra
[7] supra
[8] supra
[9] supra