When a formulation patent claims an exact numerical concentration, how far can its boundaries be stretched after the fact? The Delaware federal court addressed this question in Vertex Pharmaceuticals’ patent suit against Lupin, over a generic version of a cystic fibrosis treatment based on the active ingredient ivacaftor. Vertex’s patents claimed a concentration of “about 80%” of the active ingredient, while Lupin’s generic product contained only about 74%. The court held there was no infringement — neither literal nor under the doctrine of equivalents – and that identical bioequivalence between the products is not, on its own, enough to prove infringement of a formulation patent.
The U.S. District Court for the District of Delaware has rejected Vertex Pharmaceuticals’ patent infringement claims against Lupin in relation to Lupin’s proposed generic version of Vertex’s cystic fibrosis treatment, Kalydeco.
The decision is a useful reminder that where pharmaceutical formulation patents claim precise numerical limitations – such as a specific concentration of an active ingredient – courts may be reluctant to expand those limits after the fact. This is particularly true where the patentee relied on that precise numerical value during prosecution to overcome prior art.
As Judge Bibas put it at the outset of the decision: “Words can be twisted, but numbers do not lie.”
Vertex developed Kalydeco, a treatment for cystic fibrosis based on the active ingredient ivacaftor. The formulation challenge was that ivacaftor has poor solubility, making it difficult for the body to absorb effectively.
To address this problem, Vertex developed a formulation in which ivacaftor is dispersed with other ingredients in an amorphous solid dispersion, improving dissolution during digestion.
The patents at issue concerned formulations containing ivacaftor at specified concentrations. Two of the asserted patents claimed a drug containing exactly 80% ivacaftor, while two others claimed “about 80%” ivacaftor.
Lupin’s proposed generic product, however, contained approximately 74.257%–74.258% ivacaftor, together with different non-ivacaftor ingredients and a different manufacturing process.
The prosecution history played a central role in the court’s analysis.
Vertex had previously disclosed that an effective ivacaftor drug could include a range from “about 10% by weight to about 80% by weight” ivacaftor. During prosecution, the Patent Office initially rejected Vertex’s claims in view of Vertex’s own prior disclosures.
To obtain patent protection, Vertex argued that a formulation containing 80% ivacaftor – no more and no less – produced surprising solubility results. The examiner accepted that argument and allowed claims focused on 80% or “about 80%” ivacaftor.
However, when Vertex later tried to claim a broader range of about 72% to about 88% ivacaftor, the examiner rejected that broader range as obvious. The examiner reasoned that unexpected results shown at 80% did not necessarily extend across a wide range.
This history proved critical. The court held that Vertex was bound by the narrower claim scope it had successfully prosecuted, and could not later use infringement arguments to recapture a broader range it had failed to obtain.
Vertex did not meaningfully argue that Lupin’s 74% formulation literally infringed claims requiring exactly 80% ivacaftor. Instead, Vertex focused on the claims requiring “about 80%” ivacaftor.
Vertex argued that “about 80%” should be understood more broadly as a “high drug load,” potentially covering a range as wide as approximately 70% to 90%.
The court rejected that argument. It found that “high drug load” did not appear in the patents or in the prosecution history, and that Vertex’s expert had not provided a scientific or principled basis for treating Lupin’s 74% formulation as literally the same as “about 80%.”
By contrast, the court credited Lupin’s expert, who explained that “about” means “nearly the same as” and, in this context, would not extend beyond a small deviation from 80%. The court also noted that the patent examiner had treated “about 80%” as extending only very narrowly — for example, to 79.9%.
Because Lupin’s formulation differed from 80% by more than a few percentage points, the court found no literal infringement.
Vertex also argued that Lupin infringed under the doctrine of equivalents. The court rejected that argument as well.
The doctrine of equivalents is intended to address situations where language cannot capture every nuance of an invention. But it is not a licence to erase claim limitations, particularly where the claim language is narrow and the prosecution history confirms that the limitation was important.
Here, the court found that Vertex’s doctrine of equivalents theory would effectively eliminate the numerical limitation from the claims. Vertex had claimed 80% or “about 80%,” but was seeking to cover a materially lower concentration of approximately 74%.
The court also relied on prosecution history estoppel. Vertex had emphasized the unexpected results at 80% during prosecution, and had failed to obtain broader range claims. It could not later use the doctrine of equivalents to obtain, through litigation, the broader protection it had not secured before the Patent Office.
A particularly important part of the decision concerns Vertex’s reliance on bioequivalence.
Vertex argued that Lupin’s product was functionally similar because both products used an amorphous dispersion of ivacaftor to treat cystic fibrosis. The court rejected this as a basis for infringement.
The court explained that the function-way-result test is often ill-suited to generic drug cases, because generic drugs are expected to perform the same therapeutic function as the reference product. If bioequivalence alone were enough to prove patent infringement, the sale of virtually any generic version of a patented drug could be treated as unlawful.
The key question was not whether Lupin’s product was bioequivalent to Kalydeco, but whether it met the specific limitations of the asserted patent claims. On the facts, it did not.
The decision offers several important lessons for pharmaceutical patent owners.
First, numerical claim limitations matter. Where a patent claims a specific concentration, ratio or percentage, courts may treat that number as a real boundary of the invention.
Second, the word “about” is not automatically broad. Its scope will depend on the patent specification, the technical evidence, the understanding of a skilled person and the prosecution history.
Third, unexpected results should be supported across the full range sought. If the patentee relies on a surprising result at a single point – for example, 80% – it may be difficult later to argue that the same result justifies a much broader range.
Fourth, prosecution history can sharply limit infringement arguments. Statements made to obtain allowance may later prevent the patentee from broadening the claim through litigation.
Fifth, bioequivalence is not enough. In a formulation patent case, the fact that a generic product is therapeutically equivalent or bioequivalent to the reference product does not establish infringement unless the accused product satisfies the specific claim limitations.
For generic companies, the decision shows the potential value of designing around precise formulation claims.
A generic formulation that differs meaningfully from the claimed numerical limitations, uses different excipients, or is manufactured by a different process may have a stronger non-infringement position, particularly where the patentee’s prosecution history supports a narrow reading of the claims.
The decision also underscores the importance of expert evidence. The court placed significant weight on the credibility and care of the technical experts, and rejected an infringement theory that lacked a persuasive scientific basis.
The decision should not be read too broadly.
It does not mean that the term “about” will always be interpreted narrowly. In another case, a broader interpretation may be supported by the patent specification, technical evidence, industry practice or prosecution history.
Nor does the decision eliminate the doctrine of equivalents in pharmaceutical formulation cases. Rather, it confirms that the doctrine cannot be used to erase a numerical limitation that was important to patentability.
Finally, the decision does not weaken pharmaceutical formulation patents as a category. It simply reinforces a basic principle: the scope of protection depends on the claims that were actually obtained, not on the broader commercial protection the patentee may later wish it had secured.
The central message of the decision is that when a patent claim is framed around a numerical range, the boundaries of that range may become the boundaries of the patent protection itself. A patentee will generally face difficulty in later turning a narrow numerical point or limited range into a broader monopoly, especially where it failed to secure that broader range during prosecution. In the context of generic drugs, a meaningful numerical distance from the claimed limits may be the difference between infringement and non-infringement.
