Choosing the Ridge
The third in a five-part series of articles about how biopharma BD&L should use the patent record in a post-FOIA world.
In the Arctic, large ice sheets do not grow by extending uniformly outward. They grow by capturing frazil — small, semi-formed crystals and floating slush that drift toward the sheet’s leeward ridges and, if the conditions are right, fuse into the sheet’s mass and thicken it. The frazil, on its own, is short-lived. The sheet, without the frazil, eventually thins. Which ridge a frazil drifts into is not a passive question. It is the most consequential question the frazil’s geometry will ever face, because once the fusion occurs the frazil’s identity is the sheet’s, and the sheet’s competitors are the frazil’s.
Almost every biopharma originator working in a defended therapeutic area is, in the relevant economic sense, a frazil. The originator has a novel molecule, a scientific thesis, perhaps a Phase 2 dataset; what it does not have is a manufacturing footprint capable of supplying a global launch, distribution into thirty-plus jurisdictions, formulary access, payer contracts at scale, or a sales force calibrated to the prescriber base. The incumbent in that therapeutic area has all of those things, and built them over a fifteen-to-twenty-year horizon during which the foundational patent positions that now define the field were laid down in three or four layers. The incumbent is the ridge. Choosing — or being chosen by — a particular ridge is the partner-of-choice decision, and in a defended therapeutic area it is the only decision that matters, because the alternative is to remain a frazil and melt.
The intuitive version of this decision treats the choice of ridge as a function of fit: which incumbent’s pipeline complements the originator’s molecule, which one’s commercial reach matches the originator’s clinical positioning, which one has demonstrated willingness to do the kind of structured deal the originator’s board will accept. That version is not wrong, but it is incomplete in a specific and dangerous way. The fit analysis assumes the incumbent’s foundational position in the therapeutic area is a stable feature of the landscape, and it asks the originator only to find a ridge with the right shape. The patent record tells a different story. The incumbent’s foundational position was not given; it was assembled. It is still being assembled. And the layer of assembly currently being laid down — the patents being prosecuted right now, in the same therapeutic area, by the same incumbent and by its competitors — will determine whether the ridge the originator drifts toward will still be a ridge in five years, or whether some other ridge, currently less prominent, will have outgrown it.
The incretin field — GLP-1 receptor agonists, GLP-1/GIP dual agonists, and GLP-1/GIP/glucagon triple agonists — is a useful place to read this directly, because the foundational layers are visible by date and the current layer is visible by filing density. Novo Nordisk’s foundational position on liraglutide (Saxenda) traces to US 6,268,343, issued in July 2001. Its semaglutide position (Ozempic, Wegovy) traces to US 8,129,343, issued in March 2012. Eli Lilly’s tirzepatide foundational claim is US 9,474,780, issued in October 2016. Those three patents define the incumbents the field still organizes around. They also bracket a roughly fifteen-year arc during which the field’s commercial structure was largely dictated by which company held which foundational claim, when, and against what background of prior art. Reading those three filings does not, by itself, tell an originator which ridge to choose today. It tells the originator that the choice is being made against a record that goes back two decades and is still actively being written.
What a current full-text patent universe shows about the active writing is the part the originator’s BD&L team needs to see clearly, because it is the part that has changed since 2016 and continues to change. An analysis of 1,553 GLP-1 / dual-agonist / triple-agonist biopharma patents published from January 2016 forward, scored against the same six-dimension specification rubric this series has been describing — abstract specificity, claim-type strength, fallback hierarchy, prior-art neighborhood, Amgen v. Sanofi enablement, and survivability under Inter Partes Review (IPR) — produces a top decile of approximately 156 patents and a structural picture that does not match the press-release version of the field.
The press-release version is a two-anchor field: Lilly and Novo Nordisk, in that order or the reverse, with everything else competing for the bottom 5% in the way the Abbott diagnostics executives once described their blood-screening market — fifty smaller filers competing ferociously with each other for the residual position while the incumbents picked off the survivors. The patent record does not refute the commercial reality of the two-anchor field; Lilly and Novo run the manufacturing plants, the formulary contracts, and the prescriber base, and a frazil that fuses with either becomes part of the largest ice sheet in the field. But the patent record refuses to organize neatly around them. In the top decile of the 2016-forward universe, Sanofi holds 19 patents. Hanmi Pharm holds 18. Terns Pharmaceuticals holds 11. Pfizer holds 10. Gasherbrum Bio holds 8. Gilead holds 7. Kallyope holds 7. Indiana University holds 5. Amgen holds 4. Novo Nordisk holds 4. Eli Lilly holds 1. The mature commercial leaders of the field are not the leaders of the current foundational filing layer. The current foundational filing layer is being written substantially by adjacent incumbents (Sanofi, Pfizer, Amgen) and by what would conventionally be called the frazil cohort (Hanmi, Terns, Gasherbrum, Kallyope).
The mirror image of that picture is also legible. In the bottom decile of the same universe — patents whose specifications score weakly across the same six dimensions — Novo Nordisk holds 30, Sanofi-Aventis holds 13, MannKind holds 5, Adocia holds 4, Zealand holds 4. The bottom-decile pattern reflects, for the most part, late-stage formulation, manufacturing, delivery, and combination filings that are real and valuable defensively but do not carry foundational claim weight. The same companies are present in both deciles for different reasons; the spec rubric is reading the difference accurately.
What this means for an originator choosing a ridge is concrete. A frazil that fuses with Lilly is fusing with a company whose current top-decile foundational filing in this universe is one patent. A frazil that fuses with Novo is fusing with a company whose current top-decile presence is four patents and whose bottom-decile presence is thirty. A frazil that fuses with Sanofi — a company with no marketed GLP-1 obesity asset of its own at the scale of the two anchors — is fusing with the single largest current top-decile filer. Each of those fusions has a different geometry, and the press-release version of the field cannot distinguish them. The patent record can.
None of this argues that an originator with a Phase 2 GLP-1/GIP dual asset should reflexively partner with Sanofi rather than Lilly because Sanofi’s recent filing density is higher. The two-anchor commercial reality is real, and the originator’s board will, correctly, weigh launch capability heavily. What the patent record argues is that the choice of ridge is not the choice the press-release version of the field describes. Lilly and Novo are the largest sheets but not the densest current filers; Sanofi and the frazil cohort are the densest current filers but not the largest sheets; the layer being written now will determine which sheet thickens and which thins over the next decade. An originator that reads the patent record before choosing has a basis for a ridge selection that the qualitative version of the analysis cannot produce.
The frazil chooses. That is the part of the analogy that the older biopharma vocabulary, with its talk of being acquired and being picked off, tended to obscure. A frazil drifting toward the wrong ridge does not get a second drift.
This is the third of five articles on what biopharma BD&L practitioners can read from the patent record now that the FOIA-disclosed contract record has ended. Here’s the start of the series .
The next article looks at why, in a therapeutic area defined by a single foundational claim, filing date and claim scope determine whether a fast follower has a defensible position or a marketing budget.