The Delhi High Court's ruling in Intra-Cellular Therapies reaches a defensible result on Section 3(d), but its novelty reasoning under-engages classical genus-species doctrine, and its treatment of a co-inventor's affidavit resolves — without confronting head-on — a genuine natural-justice question about curing Controller silence on appeal rather than by remand. This companion piece examines where the judgment's reasoning holds up, and where it leaves real doctrinal gaps for future litigation.

I. Novelty and the "Coverage vs. Disclosure" Question

The core novelty dispute was a classic genus-versus-species-selection issue: the appellant argued that arriving at its claimed species from the generic Formula I of D1/D7 required "multiple selections" among independent variables, and that novelty — unlike obviousness — is ordinarily tested against a single prior document read as a whole.

One point of accuracy first: the Court did not, in fact, mosaic D1 and D7 into a single novelty attack — it matched each claim family to one anticipating document (D7 to Formula I–III claims, D1 to Formula IV claims), so the single-document rule was respected on the Court's own structure. The genuinely live question is therefore not mosaicing across documents, but whether multiple selections required within a single genus defeat its anticipatory effect.

On that question, the Court's reliance on AstraZeneca AB and Boehringer Ingelheim v. Vee Excel collapses the "covered vs. disclosed" distinction that novelty jurisprudence ordinarily maintains: a claim covers a species even if unexemplified (protecting the patentee in infringement), but whether an unnamed species falling within a prior generic formula is "disclosed" for anticipation purposes classically requires asking whether the skilled person would at once envisage that specific compound — the "photograph" test from General Tire-type genus-species jurisprudence. The judgment does not engage this framework at all; there is no reference to General Tire & Rubber Co. v. Firestone or the "individualisation" requirement novelty doctrine ordinarily demands for selection inventions.

The missing doctrinal link is the "person in the know." The respondent argued — relying on AstraZeneca — that where the inventor of the prior art and the claimed invention are the same party, anticipation is tested against a heightened "person in the know," not the ordinary skilled person. It is that standard, not the coverage doctrine alone, that justifies bypassing individualisation on these facts — yet the judgment never names it as the operative device, leaving the shortcut unexplained on its own terms.

The EPO divergence: the appellant highlighted that the EPO's own Search Opinion, examining the identical D1–D7 documents, found the claims novel under EPO's "purposive selection" doctrine (the "two lists" principle governing selection from substituent lists). Under current EPO practice, the technical-effect element — the appellant's differential-metabolism argument — bears on inventive step, not novelty, following the 2019 Guidelines revision and T 261/15; the India/EPO divergence may therefore lie as much in inventive step as in novelty. The Court's response — that patent rights are territorial — answers whether foreign grants bind the Controller, but not the separate question of why the same documents, evaluated under materially similar standards, produced opposite conclusions. A more rigorous judgment would have distinguished the EPO's framework on its merits rather than resting on non-bindingness alone.

II. Section 3(d): Correct Framework, Compressed Application

This ground is more solidly reasoned. The Court correctly extracts the Novartis triad — identify the known substance and known efficacy; explain how the claimed substance is a derivative; objectively compare therapeutic efficacy — and applies Taiho Pharmaceutical's three-factor test. Its core holding — that improved pharmacokinetic data is not, without more, "therapeutic efficacy" — faithfully tracks Novartis and Natco Pharma v. Novartis (Division Bench).

The gap: the appellant's objection that neither hearing notice identified the "known substance" (Taiho factor (i)) receives no direct answer — the Court's own analysis effectively does the Controller's work retrospectively, identifying compound Q and testing efficacy against it. This raises a natural-justice question the judgment doesn't address: if Taiho demands the Controller identify the known substance, is it appropriate for the appellate court to supply that missing step de novo rather than remand? That said, the objection's force is blunted here — compound Q was, in substance, the appellant's own admittedly-known parent molecule around which the specification was built, so the "known substance" was never genuinely in dispute, giving the omission a strong harmless-error character.

III. The Co-Inventor's Affidavit and Natural Justice

This is the judgment's most doctrinally interesting issue. The appellant's strongest procedural argument was that the Controller's order made no reference whatsoever to Dr. Peng Li's affidavit, filed three separate times, invoking Milliken & Co. v. Controller and Regents of University of California v. Union of India — both treating failure to consider evidence on record as a serious defect.

The Court's resolution turns on how a Section 117A appeal should be characterised: as a full merits rehearing, the High Court is entitled — arguably obliged — to examine evidence the Controller overlooked, which is exactly what it did, finding the affidavit's data duplicative of material already assessed. On that view, the Court's independent examination is not itself a defect; it is what a rehearing requires. Two narrower questions remain:

  1. Is reasons-giving a free-standing right? Do Milliken and Regents of the University of California treat the duty to engage with filed evidence as an independent natural-justice entitlement (remand-worthy in itself), or merely as an input into a result the appellate court may reach for itself? The judgment assumes the latter without argument — if Milliken means the former, curing the silence on appeal does not discharge the duty, because a speaking-order obligation isn't satisfied by a reviewing court supplying reasons after the fact.
  2. The constructive path not taken: a more principled disposal would have either remanded for the Controller to engage with the affidavit, or expressly articulated a harmless-error doctrine for patent prosecution explaining why first-instance failure to consider evidence is cured by appellate rehearing. Indian jurisprudence has not yet clearly developed the latter, and the judgment would have been stronger for naming and adopting it.

IV. The Inventive Step Non-Decision

Declining to examine obviousness because novelty and 3(d) were upheld is logically defensible — anticipated claims cannot, a fortiori, possess inventive step. But it remains unsatisfying in three respects: it leaves unresolved whether the Controller's "principle-plus-base-compound equals obvious" reasoning meets the motivation-to-combine bar from Hoffmann-La Roche v. Cipla and Agriboard; it provides no precedential guidance on whether that reasoning is an acceptable shortcut for future applicants facing similarly conclusory rejections; and the Court's own 3(d) analysis — using data (Example 7, the affidavit) generated to counter obviousness — arguably repeats the very conflation of novelty/obviousness with 3(d) that the appellant specifically flagged, without addressing that objection.

V. Practical Implications

For patent applicants (particularly deuterated/isotope and selection-invention filings): pharmacokinetic superiority alone will not satisfy Section 3(d) — build the specification and prosecution record around direct therapeutic/pharmacodynamic efficacy data, not exposure or metabolic-stability data alone. Do not rely on "multiple selections needed" arguments where your own earlier-filed genus patents cover the claimed variable combinations — Indian courts currently treat claim coverage as equivalent to disclosure, a stricter standard than EPO selection-invention doctrine, and same-applicant prior art may draw the heightened "person in the know" test. Foreign search opinions and grants are not persuasive authority before Indian tribunals on this evidence — cite them only as background, with independent India-specific technical argument.

For opponents and generic challengers: this is a favourable precedent for challenging deuterated-analogue and other incremental pharmaceutical patents, reinforcing that bioavailability ≠ efficacy, and providing an additional Delhi HC authority for "coverage defeats novelty" usable in oppositions or revocations against species patents built on an applicant's own earlier genus filings.

For portfolio strategy: companies holding prior Markush/genus patents face a judicially reinforced risk that subsequent deuterated or optimised species filings will be anticipated by their own earlier applications — freedom-to-operate and patentability clearances should map claim scope across the entire prior-filed family, not just third-party art.

Actionable Takeaways

  1. Build specifications around direct therapeutic/pharmacodynamic efficacy data, not PK/metabolic-stability comparisons alone.
  2. Do not rely on "multiple selections needed" novelty defences where related-party genus art covers the claimed combinations.
  3. File affidavits early and insist on express Controller engagement with them at the hearing stage — appellate curing is not guaranteed to favour the applicant, and removes an argument that would otherwise support remand.
  4. Do not treat foreign search opinions or office actions as persuasive authority on novelty/inventive step before Indian tribunals.
  5. Preserve inventive-step arguments distinctly in every proceeding — this judgment leaves the "motivation to combine" standard for chemical/pharmaceutical obviousness unaddressed on the merits.
  6. Map claim scope across your entire prior-filed patent family before filing deuterated or optimised species applications, not just against third-party art.
Practitioners should treat this decision as a useful, applicant-unfriendly precedent on Section 3(d) and novelty-by-coverage in the pharmaceutical/deuterated-compound space — while recognising its inventive-step non-decision and procedural shortcuts leave real doctrinal questions for future litigation to resolve more fully. See our companion piece for the full case analysis and procedural history.