(This article is written by Ashika Dutta, New Law College, Pune, B.B.A. LL.B., Third Year during her internship at LeDroit India)

Scope of the Article

• The meaning of pharmaceutical innovation, incremental invention and patent evergreening.
• The statutory design of Sections 2(1)(j), 2(1)(ja), 3(d), 25, 64, 83 and 84 of the Patents Act, 1970.
• The factual and procedural history of Novartis AG v. Union of India and the Supreme Court’s interpretation of enhanced efficacy.
• The distinction between therapeutic efficacy and advantageous physical properties such as stability, flow and hygroscopicity.
• Post-Novartis decisions through 2026 on the known substance, comparative data, species claims, procedural fairness and genuine incremental innovation.
• Practical lessons for patent applicants, examiners, generic manufacturers, courts and public-health policy.

Abstract

Pharmaceutical patent law must reward research without allowing nominal changes to prolong exclusivity over known medicines. India addresses this problem through Section 3(d) of the Patents Act, 1970. In Novartis AG v. Union of India, the Supreme Court refused a patent for the beta crystalline form of imatinib mesylate because the claimed advantages did not establish enhanced therapeutic efficacy over the known substance. This article explains why the judgment is neither a ban on incremental innovation nor an automatic licence for generic copying.

It situates Section 3(d) beside novelty, inventive step, disclosure, opposition and compulsory licensing; examines the evidence required to connect improved properties with therapeutic performance; and traces later decisions through 2026. The article argues that the most defensible boundary is evidence-based: a genuine new form may be patented when its therapeutic contribution is credibly demonstrated, while strategic reformulation without such contribution remains outside patentability. This approach protects meaningful innovation, procedural fairness and access to medicines together.

Keywords: Pharmaceutical Patents; Section 3(d); Evergreening; Therapeutic Efficacy; Novartis; Access to Medicines

1. Introduction: The Boundary Patent Law Must Draw

A medicine rarely moves from laboratory insight to patient use in one leap. Researchers identify a promising molecule, refine a salt or crystal form, improve stability, develop a safer dosage, discover a new delivery route and test clinical performance. Some later steps require substantial skill, cost and risk. Others may be commercially useful yet add little to the medicine’s therapeutic effect. Patent law must distinguish an invention deserving a fresh period of exclusivity from a minor alteration used mainly to postpone generic competition. That difficult line is the difference between incremental innovation and evergreening.

Evergreening is not a statutory term and should not be used as a conclusion before analysis. It generally describes strategies that extend market exclusivity through successive patents on modifications, formulations, uses or delivery systems surrounding an older product, without a commensurate inventive or therapeutic contribution. A secondary patent is not automatically abusive: a new formulation that materially improves adherence, safety or treatment outcomes may represent real innovation. The legal task is therefore not to count patents around a drug, but to test each claim against the statutory standards and the evidence.

The Supreme Court’s decision in Novartis AG v. Union of India remains the controlling Indian explanation of that boundary. The dispute concerned the beta crystalline form of imatinib mesylate, used in the anti-cancer medicine sold as Glivec or Gleevec. The Court held that the claimed form was a new form of a known substance and had not been shown to possess enhanced therapeutic efficacy. The decision became a global symbol of access to medicines, but its legal lesson is narrower and more useful: Section 3(d) does not reject all improvement patents; it demands a particular justification when the claim is directed to a new form of what is already known.

2. Innovation, Incremental Invention and Evergreening

2.1 Three concepts that must not be collapsed

Breakthrough innovation. A new active compound or platform may open a treatment pathway that did not previously exist. It still must be new, involve an inventive step and be capable of industrial application, but it normally does not encounter the first limb of Section 3(d) unless it is in truth a new form of a known substance.

Incremental innovation. A later development improves an existing medicine: a different polymorph, salt, formulation, combination, dosage form, delivery device or manufacturing process. Incremental work can solve genuine clinical and manufacturing problems. Its patentability depends on the claim, the closest prior art, the technical advance, the statutory exclusions and the evidence—not on the pejorative label “secondary patent.”

Evergreening. The concern arises when serial claims convert routine or insufficiently meaningful changes into overlapping exclusivities that delay competitive entry. Patent term is ordinarily measured from each filing, so a family of later patents can create practical barriers even after the original product patent expires. Section 3(d) is one response, but novelty, inventive step, sufficiency, opposition, revocation, competition law and compulsory licensing perform different functions.

3. India’s Statutory Framework

3.1 The ordinary patentability requirements

Under the Patents Act, 1970, an “invention” under Section 2(1)(j) is a new product or process involving an inventive step and capable of industrial application. Section 2(1)(ja) connects inventive step with technical advance, economic significance or both, and requires that the invention not be obvious to a person skilled in the art. These are separate inquiries. Novelty asks whether the claimed subject matter was previously disclosed. Inventive step asks whether the advance would have been obvious. Section 3 then identifies subject matter that is not treated as an invention even if it appears new in a literal sense.

The Delhi High Court’s Division Bench decision in F. Hoffmann-La Roche Ltd. v. Cipla Ltd. explains a structured obviousness analysis: identify the skilled person and common general knowledge, identify the inventive concept, compare it with the state of the art and decide whether the differences would be obvious without hindsight. A patent applicant cannot answer an obviousness objection merely by pointing to Section 3(d) data, and an examiner cannot reject a claim under Section 3(d) simply because it appears obvious. Each test has its own legal question.

3.2 Section 3(d): text, explanation and structure

The first limb of Section 3(d) excludes the mere discovery of a new form of a known substance when that new form does not enhance the known efficacy of the substance. The provision also excludes the mere discovery of a new property or new use for a known substance, and the mere use of a known process, machine or apparatus unless it yields a new product or employs at least one new reactant. Its Explanation treats salts, esters, ethers, polymorphs, metabolites, pure forms, particle sizes, isomers, mixtures of isomers, complexes, combinations and other derivatives as the same substance unless they differ significantly in properties concerning efficacy.

The first limb therefore requires a sequence. The decision-maker must identify the relevant known substance and its known efficacy; determine whether the claim is a new form or derivative of that substance; identify the property said to be enhanced; and decide whether the evidence shows a significant enhancement of efficacy. If the claimed compound is not a new form of the alleged known substance, Section 3(d) cannot be invoked merely because the claim belongs to the same broad chemical family. If it is a new form, ordinary commercial superiority does not by itself answer the efficacy question.

3.3 Related safeguards: opposition, revocation and public interest

Section 3(d) operates at examination and can also support pre-grant opposition under Section 25(1), post-grant opposition under Section 25(2), and revocation under Section 64. These procedures allow scientific and legal scrutiny at different stages. The Delhi High Court’s 2024 decision in Novartis AG v. Natco Pharma Ltd. stressed that a statutorily conferred pre-grant hearing cannot be treated as expendable or cured casually by a later remedy. Anti-evergreening policy is not a licence to dilute procedural fairness; reliable patent quality depends on both correct standards and a fair opportunity to answer objections.

Sections 83 and 84 serve a different purpose. Section 83 states general principles about working patents, technological innovation, public interest and reasonable affordability. Section 84 permits compulsory licensing after the prescribed period where public requirements are unmet, the invention is not available at a reasonably affordable price, or it is not worked in India. In Bayer Corporation v. Union of India the compulsory-licence litigation concerning sorafenib illustrated that a valid patent can still be subject to access-oriented remedies. Patentability and compulsory licensing must not be confused: Section 3(d) asks whether a qualifying patent should exist; Section 84 regulates use of an existing patent in specified circumstances.

4. Novartis AG v. Union of India: From Laboratory Form to Legal Test

4.1 The claimed invention and the “mailbox” setting

Novartis traced its work from imatinib in free-base form to imatinib mesylate and then to the beta crystalline form of that salt. The earlier Zimmermann patent disclosed imatinib and derivatives and was central to the prior-art analysis. In 1998 Novartis filed the Indian application for the beta crystalline form during India’s transitional “mailbox” period, when pharmaceutical product applications could be filed but were to be examined after India introduced full product-patent protection in 2005. The application described useful properties of the beta form, including better flow, thermodynamic stability and lower hygroscopicity.

Following examination and oppositions, the Assistant Controller rejected the application in 2006. The Intellectual Property Appellate Board accepted that the claim satisfied certain ordinary patentability considerations but held it barred by Section 3(d) for want of enhanced efficacy. In separate writ proceedings, Novartis challenged the amended provision. The Madras High Court decision of 6 August 2007 rejected the Article 14 challenge and held that treaty-compliance complaints belonged to the dispute-settlement framework rather than a domestic writ adjudication of TRIPS. The patentability appeal eventually reached the Supreme Court.

4.2 What was the “known substance”?

The identification of the comparator was decisive. Novartis argued that the beta crystalline form should be compared with imatinib free base and relied on the path by which it had developed the salt and crystal. The Court examined the Zimmermann patent, scientific publications, regulatory material and the applicant’s own documents. It concluded that imatinib mesylate was a known substance with known efficacy and that the beta crystalline form was a new form of it. Once that relationship was established, the claim entered the substantive screen and Explanation to Section 3(d).

This reasoning guards against two opposite errors. A generic disclosure of a vast chemical genus should not automatically make every later species a “known substance” without careful analysis. But an applicant cannot avoid Section 3(d) merely by selecting an artificially remote comparator if the immediate substance and its efficacy were already known. Later cases have developed this point by requiring the Patent Office to identify the comparator clearly, rather than leaving the applicant to guess which prior substance and which efficacy must be answered.

4.3 Why improved physical properties were insufficient

For a medicine, the Supreme Court interpreted “efficacy” as therapeutic efficacy: the capacity to produce the desired therapeutic effect. The Court accepted that properties such as improved stability, flow and reduced hygroscopicity could be advantageous. It did not hold those properties scientifically irrelevant. It held that they did not, without a demonstrated link, establish enhanced therapeutic efficacy. A more stable crystal can simplify storage or manufacturing yet leave the medicine’s therapeutic performance unchanged. The statutory question is not whether the new form is useful, but whether the relevant efficacy is significantly enhanced.

Novartis also relied on approximately thirty per cent greater bioavailability compared with imatinib free base. The Court declined to equate increased bioavailability automatically with enhanced therapeutic efficacy. Bioavailability can be relevant evidence, but its legal significance depends on what the change does to treatment: dose, clinical response, toxicity, safety, patient outcome or another therapeutically meaningful measure. Because the material did not establish the necessary therapeutic connection against the proper known substance, the beta form failed Section 3(d).

4.4 The holding—and its deliberate limit

The Supreme Court dismissed the appeal and upheld refusal of the patent. Importantly, it warned against reading the judgment as a bar on all incremental inventions in chemical and pharmaceutical fields. The Court’s point was that this particular product did not satisfy this particular statutory test. That reservation is central. Section 3(d) is a calibrated filter: it allows a new form to pass when significant efficacy-related difference is proved, while blocking a mere change of form without that evidence.

5. Evidence After Novartis: From Assertion to Objective Comparison

5.1 A comparison, not a slogan

An applicant facing Section 3(d) should state the known substance, the known efficacy, the claimed difference, the test method and the comparative result. Absolute data about the new form may be impressive yet legally uninformative if there is no appropriate baseline. Conversely, an objection that says only “new form of a known substance” is procedurally deficient because the applicant cannot know what comparator must be addressed. The inquiry must be objective and reproducible enough for an examiner, opponent and court to understand the claimed enhancement.

In D.S. Biopharma Ltd. v. Controller of Patents the Delhi High Court required the Patent Office to identify the specific known substance and explain how the claimed compound is its new form. Ischemix LLC v. Controller of Patents similarly emphasised a fair chance to respond and discussed the conditions under which post-filing material may support an efficacy assertion already grounded in the specification. These decisions improve both rigor and fairness: the applicant bears the burden of proving enhancement, but the objection must disclose the case to be met.

5.2 Can post-filing data be considered?

Pharmaceutical evidence often develops after a patent application is filed, while patent law also insists that the application disclose the invention rather than reserve its essence for later proof. A sensible distinction is between later evidence that confirms or elaborates a technical effect plausibly disclosed in the specification, and later material that invents an entirely new effect absent from the filing. The former may assist evaluation; the latter risks replacing disclosure with hindsight. Applicants should therefore include comparative design, endpoints and available results at filing, and not assume that a later affidavit can cure a silent specification.

5.3 Therapeutic efficacy is product-sensitive

Novartis addressed a pharmaceutical compound. The meaning of efficacy must relate to the function of the substance under examination. In Novozymes v. Assistant Controller of Patents the Madras High Court considered phytase variants used in animal feed and accepted that efficacy could be assessed by the function of the biochemical product rather than importing a cancer-drug clinical test mechanically. The broader lesson is contextual: identify what the known substance is supposed to do, then test whether the new form significantly improves that function.

6. Later Decisions: Refining the Boundary

6.1 A species is not automatically a new form of a known substance

In Kudos Pharmaceuticals Ltd. v. Natco Pharma Ltd. the Delhi High Court addressed olaparib in patent litigation. The Court rejected a Section 3(d) contention at the interim stage because the alleged earlier genus disclosure did not make olaparib itself a “known substance” on the facts analysed. The decision is important because it prevents Section 3(d) from swallowing the rules of anticipation: the fact that a species falls within a broad Markush formula does not invariably establish that the species was known. Specific disclosure, selection and the teaching of prior art still matter.

6.2 The Patent Office must identify its case

In Taiho Pharmaceutical Co. Ltd. v. Controller of Patents the Delhi High Court restated three elements needed for a sustainable Section 3(d) objection: identification of the known substance with known efficacy; an explanation of how the claimed substance is a derivative or new form; and an objective comparison of therapeutic efficacy. The decision does not shift the ultimate burden away from the applicant. It ensures that the burden is responsive to a legally identified comparator rather than an unspecified allegation.

The same concern appears in Osaka University v. Assistant Controller of Patents decided on 24 December 2025. The court remanded the matter because the relevant known compound and prior art had not been specified in the examination report and hearing notice, while substantive reasoning emerged only in the refusal order. Patent quality is not improved by surprise. A reasoned notice allows the applicant to supply targeted evidence and produces a record that can withstand review.

6.3 Recent application in 2026

In Intra-Cellular Therapies, Inc. v. Controller of Patents delivered on 6 July 2026, the Delhi High Court considered claimed deuterated compounds against identified prior compounds and separate objections of novelty, inventive step and Section 3(d). The judgment examined pharmacokinetic and metabolism data but held that the material did not establish the required enhanced therapeutic efficacy and upheld the Section 3(d) objection. It also demonstrates that foreign grants do not control Indian patentability: patent rights are territorial and the Indian statutory standards must be independently satisfied.

Read together, these authorities produce a balanced doctrine. Section 3(d) must not be used as an unparticularised rejection or as a substitute for novelty. Yet once the known substance and relationship are properly established, advantageous laboratory or pharmacokinetic properties must be connected by credible evidence to enhanced efficacy. The modern cases refine procedure and comparator selection without diluting the substantive demand articulated in Novartis.

7. Section 3(d) and International Patent Obligations

Article 27.1 of the TRIPS Agreement requires patents to be available for inventions in all fields of technology when they are new, involve an inventive step and are capable of industrial application, subject to permitted exclusions and conditions. TRIPS does not supply exhaustive definitions of “invention,” “inventive step” or the precise threshold for secondary pharmaceutical claims. Members retain room to formulate patentability standards within the agreement. India uses Section 3(d) to define when certain discoveries of new forms do not count as inventions.

The WTO’s account of the Doha Declaration on TRIPS and Public Health records the shared understanding that TRIPS should be interpreted and implemented in a manner supportive of public health and access to medicines, while preserving patent protection and its flexibilities. Section 3(d) should not be defended by saying access always defeats innovation, nor attacked by saying every commercially useful modification deserves a patent. Its legitimacy lies in a defensible patentability threshold applied without discrimination and with scientific evidence.

8. Practical Comparison: Genuine Improvement or Evergreening Risk?

9. Illustrations

Illustration A: a more stable polymorph

A company discovers a polymorph of a known drug that tolerates humid storage and flows better during tablet manufacture. Those properties may reduce waste and make production easier, but they do not by themselves show that patients obtain a better therapeutic result. If no significant efficacy-related difference is demonstrated, the new form risks exclusion under Section 3(d), even if a new manufacturing process could separately qualify on its own facts.

Illustration B: improved exposure linked to safer treatment

A new salt of a known medicine produces consistent exposure at half the dose and comparative studies show a materially lower rate of a dose-related serious adverse effect while maintaining clinical response. Bioavailability alone would not automatically satisfy Section 3(d), but the demonstrated relationship among exposure, dose, safety and treatment performance provides a stronger therapeutic-efficacy case. Novelty, inventive step, sufficiency and claim scope must still be proved.

Illustration C: a genuinely selected species

A later researcher selects one compound from a very broad earlier genus. The earlier document does not specifically disclose it, provide a path to select it or reveal its properties. The compound shows an unexpected therapeutic profile. The correct analysis first asks whether the species is actually known and obvious. Section 3(d) should not be triggered merely because the compound can be drawn somewhere inside a vast generic formula; if the species is not a new form of a known substance, the enhanced-efficacy screen is not the starting point.

10. Lessons for Stakeholders

10.1 For patent applicants

Applicants should design Section 3(d) strategy before filing. The specification should identify the closest known substance, explain the claimed form, disclose comparative methods and connect improved properties with the relevant efficacy. Claims should be proportionate to tested embodiments. Separate arguments should address novelty, inventive step and statutory exclusion. If later evidence is necessary, it should confirm a foundation already present rather than supply the invention for the first time. A prosecution record built this way is more persuasive than a stack of foreign grants or general statements about research cost.

10.2 For the Patent Office and opponents

Examiners and opponents should identify the exact known substance, its known efficacy, the alleged new-form relationship and the evidentiary deficiency. Genus disclosure, obviousness and Section 3(d) must not be merged into one objection. Orders should engage with submitted experiments, explain why a comparator is appropriate and distinguish a physical benefit from therapeutic enhancement. Clear notice is not procedural generosity; it is necessary for scientifically reliable adjudication.

10.3 For courts and policy-makers

Courts should preserve the narrow, evidence-based character of Novartis. An excessively lax standard would permit strategic repetition, while an inflexible demand for completed large-scale clinical trials at filing could make legitimate improvement patents practically impossible. The law can evaluate animal data, in vitro results, pharmacokinetics, safety and clinically relevant surrogates according to the claimed effect and the maturity of the science, while insisting on a reasoned therapeutic connection. Policy should also distinguish patent validity from price regulation, procurement, competition law and compulsory licensing, each of which addresses a different problem.

11. A Five-Step Working Test

• Identify the claim precisely: product, form, derivative, use, formulation, combination or process.
• Apply ordinary patentability: determine novelty, inventive step and industrial applicability without hindsight.
• Trigger Section 3(d) only after identifying a specific known substance and explaining the new-form relationship.
• Compare efficacy objectively: define the relevant function, comparator, endpoint, method, magnitude and therapeutic connection.
• Calibrate the result: grant only claim scope supported by the demonstrated contribution; preserve opposition, revocation and access safeguards where applicable.

12. Conclusion

Novartis is best understood as a discipline of proof. It refuses to treat every change in pharmaceutical form as either worthless or patentable. Instead, it asks whether a claimed new form of a known substance makes a significant efficacy-related contribution and whether the applicant has demonstrated that contribution against the proper comparator. Physical advantages and bioavailability may matter, but their legal force depends on a credible link to therapeutic performance.

Later decisions have made the doctrine more precise. The Patent Office must identify the known substance and give fair notice; a broad genus does not invariably make a species known; post-filing evidence cannot substitute for a missing technical foundation; and novelty, inventive step and Section 3(d) remain distinct. At the same time, recent adjudication confirms that foreign grants and advantageous pharmacokinetics do not displace India’s independent requirement of enhanced efficacy.

The enduring lesson is therefore not innovation versus access, as if one must defeat the other. Patent law supports both when it rewards an evidence-backed therapeutic advance and withholds a fresh monopoly from a merely cosmetic extension. Section 3(d), applied carefully, gives genuine incremental innovation a path to protection while keeping evergreening outside the gate.