On May 13, 2026, the Federal Circuit closed out a patent fight that had already run for more than six years, and it did so on grounds that had nothing to do with the accused generic drug’s ingredients. In Actelion Pharmaceuticals Ltd. v. Mylan Pharmaceuticals Inc., the court affirmed that Mylan’s generic version of Actelion’s pulmonary hypertension drug Veletri did not infringe, not because the chemistry was different, but because Actelion had narrowed its own patent claim during prosecution and then tried to recapture that narrowed ground years later in litigation [1]. The case had already produced one stipulated judgment in Actelion’s favor, based on a claim construction that was later vacated. It took a second trip through claim construction, a bench trial, and a full read of the prosecution history to reach the opposite result [1].
That sequence is the subject of this article. Settlement strategy in Hatch-Waxman litigation is usually priced early, often before anyone on either side has done a close read of the file wrapper, the complete record of communications between a patent applicant and the U.S. Patent and Trademark Office. Once that record gets read closely, and it eventually does, on undamaged patents it changes very little. On patents with a narrowing amendment, an unexplained examiner’s edit, or a disclosed-but-unclaimed alternative sitting in the specification, it can change the settlement number entirely.
The Short Answer
Prosecution history estoppel bars a patent owner from using the doctrine of equivalents to recapture claim scope it gave up during patent examination, typically through a narrowing amendment made to overcome a rejection [3][4]. In ANDA litigation, generic challengers and their counsel now routinely pull the full prosecution history before settlement talks mature, because a defensible file wrapper changes what a case is worth. Two recent Federal Circuit decisions, Actelion v. Mylan (2026) and Amgen v. Amneal (2020), show what happens on each side of that line: one brand patent survived years of litigation on a claim construction that ultimately failed once the prosecution history was fully weighed, and one brand patent’s equivalents theory was barred outright because an unexplained examiner’s amendment had narrowed 272 possible binder-disintegrant combinations down to 12 [1][2].
What Hatch-Waxman Settlement Actually Looks Like Before Discovery
The Baseline Settlement Rate in Paragraph IV Cases
Most Paragraph IV patent disputes never reach a verdict. Estimates of the settlement rate vary by dataset and time period. An academic study of paragraph IV ANDA filings litigated to a district court decision between 1988 and 2012 found that, among cases involving active-ingredient patents where the ANDA was the first filed for that ingredient, a majority of filed lawsuits reached a decision rather than settling, a litigation rate the authors describe as unusually high compared with patent litigation generally, where 90 percent or more of filed suits settle [6]. A separate dataset covering patents listed in Paragraph IV notice letters found that of 421 relevant patents studied, 72 were approved without a court challenge, 159 were litigated and then settled, and 192 were litigated to a decision [6]. That works out to roughly 45 percent of litigated active-ingredient-patent cases settling rather than going to judgment, a lower settlement rate than patent litigation carries on average.
Separately, the number of final Hatch-Waxman patent settlements filed with the FTC and Department of Justice under the Medicare Prescription Drug, Improvement, and Modernization Act runs in the low-to-mid 200s annually in recent reporting years: 232 in fiscal year 2016 and 226 in fiscal year 2017, close to the prior record [7][8]. Those figures count all filed settlements, not just first-to-file or active-ingredient disputes, so they are not directly comparable to the academic litigation-to-decision study above, but they establish that settlement volume in this category is large and consistent year over year.
Why Early Settlement Numbers Get Set Before Anyone Reads the File Wrapper
A brand manufacturer’s opening settlement position, and a generic challenger’s willingness to accept it, both get set early in the case, usually before claim construction and often before either side’s litigation counsel has done a line-by-line review of the prosecution history. That early number is built on the patent’s face: the claims as issued, the drug’s remaining market exclusivity, and a rough sense of the infringement theory. The file wrapper sits in the USPTO’s Patent Application Information Retrieval system the entire time, available to anyone who wants it, but reading a multi-year prosecution history closely, sometimes running to hundreds of pages across several continuation and divisional applications, is not something litigation teams do as a first step. It is done once a case looks like it is worth the cost of doing it.
What Prosecution History Estoppel Is
The Doctrine of Equivalents in One Paragraph
Patent infringement can be proven two ways: literally, when the accused product meets every claim limitation exactly, or under the doctrine of equivalents, when the accused product differs from the claim in a way that is insubstantial, performing substantially the same function in substantially the same way to achieve substantially the same result. The doctrine of equivalents exists so that a patent’s protection is not defeated by trivial design-arounds. Prosecution history estoppel is the doctrine’s counterweight: it prevents a patent owner from using the doctrine of equivalents to reclaim territory that the owner gave up during prosecution in order to get the patent granted in the first place.
Warner-Jenkinson: The Presumption That a Narrowing Amendment Surrenders Ground
In Warner-Jenkinson Co. v. Hilton Davis Chemical Co., the Supreme Court held that when a patent applicant narrows a claim through amendment and does not explain why, courts should presume the amendment was made for a reason related to patentability, which triggers estoppel as to the surrendered territory [4]. The patent owner carries the burden of overcoming that presumption. Silence in the file wrapper, in other words, does not protect a patent owner. It works against one.
Festo v. SMC: The Rebuttable Presumption and Its Three Exceptions
The Supreme Court refined the doctrine further in Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., rejecting the Federal Circuit’s earlier rule that a narrowing amendment created a complete bar on equivalents. Instead, the Court adopted a rebuttable presumption: a narrowing amendment presumptively surrenders the territory between the original and amended claim language, but the patent owner can rebut that presumption in three specific ways [3].
Exception One: Unforeseeability
The patent owner can show that the accused equivalent would have been unforeseeable to a person skilled in the art at the time of the amendment, meaning there was no way to have drafted around it [3].
Exception Two: Tangential Relation
The patent owner can show that the reason for the amendment had no more than a tangential relationship to the accused equivalent, so the amendment was not really surrendering that particular ground even though it narrowed the claim [3][5].
Exception Three: Some Other Reason
The patent owner can show some other reason why it could not reasonably have been expected to describe the equivalent at the time it made the amendment [3].
All three exceptions require the patent owner to prove something from the prosecution record itself. There is no exception for simply not having thought about it, and the exceptions are evaluated by a judge, not a jury, based on the written prosecution history [3].
Case Study: Actelion v. Mylan and the Epoprostenol Patent That Took Six Years to Read Correctly
The First Judgment: A Stipulation Built on an Incomplete Construction
Actelion’s two patents on Veletri, a lyophilized formulation of epoprostenol, required that the drug be formed from a bulk solution with a pH of 13 or higher. Mylan filed an ANDA for a generic version before Actelion’s patents expired, and Actelion sued for infringement, arguing that Mylan’s product was covered literally or, in the alternative, under the doctrine of equivalents [1]. The district court’s first claim construction read “a pH of 13 or higher” as subject to ordinary rounding, meaning the claim reached down to pH 12.5. On that reading, the parties stipulated to a judgment of infringement and a permanent injunction, and Mylan appealed [1].
What Changed on Remand
The Federal Circuit vacated that judgment in 2023, holding that the claim term could not be properly construed without extrinsic evidence, and sent the case back down [1]. On remand, the district court considered both the intrinsic record, including the prosecution history, and extrinsic evidence such as United States Pharmacopeia standards, and arrived at a materially different construction: “a pH of 12.98 or higher.” Under that reading, the central factual question at a February 2024 bench trial became whether pH should be measured at the drug’s cold storage temperature or at the pharmaceutical industry’s standard measurement temperature of 25±2°C. Mylan’s bulk solution measured well below 12.98 at standard temperature. The court found that a skilled artisan would understand the claim to require measurement at standard temperature, which meant no literal infringement [1].
Two Independent Bars: Prosecution History Estoppel and Disclosure-Dedication
Actelion’s fallback argument, infringement under the doctrine of equivalents, ran into two separate doctrines, either one of which was sufficient on its own to defeat the claim. During prosecution, Actelion had amended the pH limitation more than once, with the final narrowing amendment moving from “a pH of greater than 12” to “a pH of 13 or higher” after an obviousness rejection. The prosecution record showed the examiner treated pH 13 as the threshold at which Actelion had demonstrated unexpected results, which the Federal Circuit found to be the discernible reason for the amendment. That triggered prosecution history estoppel [1]. Separately, Actelion’s own patent specification disclosed pH ranges below 13, including 12.5 to 13 and greater than 12, without claiming them, which the disclosure-dedication rule treats as a dedication of that unclaimed subject matter to the public. The Federal Circuit rejected Actelion’s argument that overlapping, non-exclusive ranges should escape the rule [1].
Case Study: Amgen v. Amneal and the Arithmetic of a Narrowing Amendment
The Examiner’s Amendment Amgen Didn’t Explain
Amgen’s U.S. Patent No. 9,375,405 covers a rapid-dissolution formulation of cinacalcet hydrochloride, the active ingredient in Sensipar, a treatment for secondary hyperparathyroidism and hypercalcemia in dialysis and parathyroid cancer patients [2]. Amgen sued twelve sets of ANDA filers for infringement; three, Amneal, Piramal, and Zydus, reached the Federal Circuit on appeal [2]. During prosecution, after Amgen amended the claimed amount of cinacalcet HCl to overcome an obviousness rejection over prior art, the examiner proposed and Amgen accepted an Examiner’s Amendment converting the claim’s binder and disintegrant elements into closed “consisting of” Markush groups. The Notice of Allowance followed shortly after. Eight months later, in an unrelated filing, Amgen’s counsel added boilerplate language stating the earlier changes were not made for patentability reasons but merely to “place the claims in proper format” [2].
From 272 Combinations to 12
Piramal’s generic product used pregelatinized starch as a binder, which Amgen argued was equivalent to one of the binders Amgen had actually claimed, povidone. The Federal Circuit disagreed that Amgen could reach that equivalent. Piramal showed that the Examiner’s Amendment had narrowed the binder-disintegrant combinations disclosed in the two cited prior-art references, Creekmore and Hsu, from 152 and 120 possible combinations, 272 combined, down to 12 [2]. The court found that reduction was too substantial to have served a merely cosmetic purpose, and it dismissed Amgen’s eight-months-later boilerplate statement as unclear, conventional language that could not carry the burden of showing the amendment was unrelated to patentability [2]. Because Amgen could not rebut the presumption, it was estopped from asserting the doctrine of equivalents against Piramal’s product, and the district court’s finding of no infringement was affirmed [2].
Three Generic Defendants, Three Different Outcomes
The same patent, litigated against three ANDA filers in the same case, produced three different results, which is itself the clearest illustration of how fact-specific this doctrine is. Amneal’s product used the binder Opadry, and the case was vacated and remanded on a separate claim-construction question about whether the claim’s binder and disintegrant limitations were open or closed to unlisted ingredients. Piramal’s noninfringement finding was affirmed on prosecution history estoppel grounds. Zydus, which used a diluent falling within the literal scope of the claimed disintegrant limitation, was found to infringe and had its judgment affirmed [2].
Prosecution history estoppel and disclosure-dedication rule (independent bars)
Narrowing amendment from “pH greater than 12” to “pH of 13 or higher”; unclaimed pH ranges disclosed in the specification
No infringement, literal or equivalents, affirmed after remand
Amgen, Inc. v. Amneal Pharmaceuticals, LLC, 945 F.3d 1368 (Fed. Cir. 2020)
Sensipar (cinacalcet HCl)
Prosecution history estoppel
Unexplained Examiner’s Amendment narrowing binder/disintegrant Markush groups from 272 combinations to 12
No infringement by Piramal affirmed; Amneal claim construction vacated and remanded; Zydus infringement affirmed
The Disclosure-Dedication Rule: A Second File-Wrapper Trap
Prosecution history estoppel is not the only doctrine hiding in a patent’s own paperwork. The disclosure-dedication rule bars a patentee from using the doctrine of equivalents to capture subject matter that the patent’s own specification discloses but does not claim. The logic is straightforward: if the patentee described something in the written description and chose not to put it in a claim, the public is entitled to treat that disclosed-but-unclaimed material as dedicated to public use. Actelion v. Mylan shows this rule operating independently of prosecution history estoppel, not as a backup argument but as a separate, self-sufficient bar. Actelion’s specification disclosed pH ranges of 12.5 to 13 and greater than 12 as alternatives to the claimed “13 or higher,” and the Federal Circuit held that disclosing overlapping, non-mutually-exclusive alternatives does not save a patentee from the rule [1]. A brand company’s own specification, written to be thorough and to support multiple embodiments, can end up supplying the generic challenger’s best noninfringement argument.
What a File Wrapper Costs to Read, and When Litigants Actually Read It
AIPLA Cost Data Behind the Rush to Settle Early
The American Intellectual Property Law Association’s biennial Economic Survey is the industry’s standard benchmark for litigation cost, broken out by the dollar amount at risk in a case. For pharmaceutical patent cases with more than $25 million at risk, which describes most branded blockbuster disputes, the median cost to litigate through trial and appeal runs around $4 million per side, with total spend in contested cases sometimes reaching $10 to 15 million once expert witnesses, discovery, and appeals are included [10][13]. Even the smallest Hatch-Waxman cases are not cheap: ANDA litigation with less than $1 million officially at risk still carries a median total cost around $900,000 [13].
“Brand-firm stakes in such cases average $4.3 billion while generic-firm stakes average $204.3 million” [6].
That figure comes from an event-study analysis of paragraph IV cases that were litigated to a district court decision, and it captures the asymmetry that shapes settlement behavior in this category: the branded manufacturer is typically defending a monopoly worth many times what the generic challenger stands to gain, which gives the brand side a strong incentive to settle on favorable terms rather than risk a full loss, and gives the generic side leverage even when its own individual stakes are comparatively modest [6].
The Point in Litigation Where the File Wrapper Gets Read Closely
Given costs in the millions of dollars, most litigation teams do not commission an exhaustive, claim-by-claim prosecution history analysis on day one. That work tends to concentrate around claim construction briefing, the point in an ANDA case where the parties are required to put their reading of every disputed claim term, and the intrinsic evidence supporting it, in front of the court. Intrinsic evidence is exactly the specification, the claims, and the prosecution history, so claim construction is where prosecution history estoppel arguments typically surface for the first time in a formal way, well after a case has been filed and often after initial settlement conversations have already started. A defensible early settlement estimate that does not account for what a file wrapper actually says is a number that can move substantially once claim construction briefing forces a real look at it, which is exactly what happened to the parties’ first stipulated judgment in Actelion v. Mylan [1].
A Taxonomy of File-Wrapper Exposure in ANDA Patents
The categories below are an original classification built from the two case studies above and general prosecution practice. They are not an established doctrinal taxonomy recognized by the courts, but a way of organizing the recurring patterns that create prosecution-history risk in pharmaceutical patents.
Type 1: The Explained Narrowing Amendment
A claim is narrowed in direct, documented response to a prior-art rejection, and the file wrapper states the reason clearly. This is the cleanest case for estoppel; there is little room to argue the amendment was unrelated to patentability, as in the pH amendment in Actelion v. Mylan [1].
Type 2: The Unexplained Examiner’s Amendment
The examiner, not the applicant, proposes a narrowing change during an interview, and the applicant simply accepts it. Because the amendment did not originate with the patentee, there can be a temptation to treat it as administrative rather than substantive. Amgen v. Amneal shows that courts will look at what the amendment actually did to claim scope, not who drafted it, and a sparse Examiner’s Amendment summary offers little protection [2].
Type 3: The Disclosed-But-Unclaimed Alternative
The specification describes ranges, formulations, or embodiments that never make it into a claim. This does not require a narrowing amendment at all; it is purely a drafting choice, and it triggers the disclosure-dedication rule rather than prosecution history estoppel, as Actelion’s unclaimed pH ranges did [1].
Type 4: The Closed Markush Group
A claim element is drafted or amended using “consisting of” language listing specific members, rather than open, “comprising”-style language. Once a Markush group is closed, equivalents outside the listed group are generally unavailable, which is what limited Amgen’s binder and disintegrant claims regardless of the estoppel analysis [2].
Type 5: The Late, Boilerplate Clarification
An applicant tries to characterize an earlier amendment’s purpose well after the fact, often in a routine filing unrelated to the amendment itself. Courts have treated generic, conclusory language added later as insufficient to rebut the presumption that a narrowing amendment was made for reasons of patentability, as with Amgen’s eight-months-later statement [2].
How New Information Moves the Settlement Number
What Happens When the File Wrapper Favors the Generic Challenger
When a close read of the prosecution history reveals a Type 1 or Type 2 exposure, a generic challenger’s noninfringement position gets meaningfully stronger, and the settlement calculus shifts with it. A brand manufacturer facing a credible estoppel argument has less incentive to hold out for a settlement date close to natural patent expiration, and more incentive to accept an earlier generic entry date, license terms, or a lower payment, because the alternative is a judgment of noninfringement with no compensation at all. This dynamic mirrors what researchers have found in the adjacent context of inter partes review: new information about a patent’s validity, generated through the PTAB process rather than prosecution history review, has a large, measurable, positive effect on the rate at which parallel district court litigation settles, precisely because it changes what both sides think the case is worth [9]. Prosecution history review works on the same principle in a different forum: it is a source of new information that recalibrates expected outcomes before a case reaches a jury or a bench trial.
What Happens When It Favors the Brand
The reverse is equally real. A brand patent with a clean prosecution history, no unexplained amendments, no disclosed-but-unclaimed alternatives, and claims that were never meaningfully narrowed to get past an examiner, gives the generic challenger less to work with under the doctrine of equivalents. In those cases, a close read of the file wrapper confirms rather than upends the early settlement estimate, and the case proceeds on the terms both sides expected. The point is not that reading the prosecution history always favors one side; it is that settlement numbers set before that review happens are provisional, whichever direction they eventually move.
What This Means for Brand Manufacturers
Prosecution strategy is litigation strategy conducted years in advance, under less scrutiny and often by different counsel than the litigators who will eventually defend the patent in an ANDA suit. Every amendment made to overcome a rejection, and every embodiment described but not claimed, becomes part of a public record that a well-resourced generic challenger’s litigation team will eventually read in full. Brand IP teams that treat the file wrapper as a future trial exhibit, rather than a closed administrative matter once the patent issues, tend to produce patents that hold their settlement value. That means documenting the actual reason for narrowing amendments on the record, scrutinizing Examiner’s Amendments before accepting them rather than treating an interview summary as a formality, and being deliberate about what gets described in the specification but left out of the claims.
What This Means for Generic Challengers
A generic company evaluating whether to file a Paragraph IV certification, and how aggressively to litigate rather than settle, gets a real return on pulling the complete file wrapper early, not after settlement talks have already anchored around a number based on the issued claims alone. The full prosecution history for any issued U.S. patent, including every amendment, examiner interview summary, and preliminary amendment, is available without cost through the USPTO’s Patent Application Information Retrieval system. A brand patent with a Type 1 or Type 2 exposure, in the taxonomy above, is a patent where a noninfringement theory under the doctrine of equivalents may already be available regardless of how the accused product is ultimately designed, which changes both the litigation budget a generic filer should expect to spend and the settlement terms it should be willing to accept.
The Financial Stakes Behind the Settlement Calculus
Average Stakes in Litigated Paragraph IV Cases
The asymmetry between brand and generic stakes described above, $4.3 billion average brand exposure against $204.3 million average generic exposure in cases that reach a decision, is not incidental to how prosecution history estoppel gets used in negotiation [6]. Because a single adverse ruling carries outsized consequences for the brand side, brand manufacturers have a strong incentive to settle rather than risk a full noninfringement judgment once a credible estoppel argument surfaces, even when that argument might ultimately fail if fully litigated. Generic challengers, facing much smaller individual stakes but the prospect of a 180-day exclusivity period if they prevail, have their own incentive to press an estoppel argument hard in negotiation rather than wait for a judge to rule on it.
How Schering-Plough v. FTC Changed the Settlement Math After 2002
The same academic study found a structural shift in settlement behavior tied to a specific legal event: after the Eleventh Circuit’s 2005 decision upholding the FTC’s earlier acceptance of a reverse-payment settlement in Schering-Plough v. FTC, settlement became significantly more likely in paragraph IV cases, and the stakes involved in settled cases dropped, even though the drugs involved had higher average brand sales than before [6]. That finding matters here because it shows settlement rates in this category are sensitive to the legal environment surrounding settlements generally, not just to the facts of an individual patent’s prosecution history. A generic challenger’s estoppel argument operates inside that broader environment; it is one input into a settlement calculus that is already shaped by antitrust exposure, exclusivity value, and the relative size of what each side stands to lose.
How Often Do Paragraph IV Cases Actually Settle? The Data
FTC’s Annual Hatch-Waxman Settlement Snapshot
The FTC has published an annual snapshot of final Hatch-Waxman patent settlements since FTC v. Actavis established that reverse-payment settlements can violate antitrust law. The two most detailed recent counts available are as follows.
FY 2016
232 final patent settlements were filed with the FTC and Department of Justice. Only one agreement contained a side deal or no-authorized-generic commitment, the lowest count since 2004, and 29 of 30 settlements with explicit compensation to the generic company involved $7 million or less in litigation-fee reimbursement, the category the Supreme Court in Actavis suggested might represent a legitimate, non-anticompetitive payment [7].
FY 2017
226 final patent settlements were filed, close to the FY 2016 total. For the first time since FY 2004, no settlement agreement that year contained a no-authorized-generic commitment, and agreements with “possible compensation” to the generic company, provisions that might function as compensation depending on marketplace circumstances, fell from 14 the prior year to 11 [8].
What Determines Whether a Specific Case Settles
A 2023 study in the International Review of Law and Economics examined Paragraph IV settlements filed between 2003 and 2020 and found that the number of patents being litigated and the years of remaining market exclusivity were the strongest predictors of whether a given case settled, while whether the litigated patent covered the active ingredient itself, as opposed to a secondary formulation or method patent, was not a statistically significant predictor [9]. That is a useful corrective to an intuitive assumption: a weaker-looking secondary patent is not automatically more likely to settle than a strong-looking primary one. Case-specific factors, including what the prosecution history actually shows, carry more weight than a patent’s category.
Methodology
This article draws primarily on two recent, precedential Federal Circuit decisions applying prosecution history estoppel and the disclosure-dedication rule in ANDA litigation: Actelion Pharmaceuticals Ltd. v. Mylan Pharmaceuticals Inc., No. 24-1641 (Fed. Cir. May 13, 2026), sourced through a Crowell & Moring client alert summarizing the opinion, and Amgen, Inc. v. Amneal Pharmaceuticals, LLC, 945 F.3d 1368 (Fed. Cir. 2020), sourced through a Harvard Journal of Law & Technology Digest case comment that cites the opinion’s pinpoint pages directly. Settlement-rate and settlement-predictor data come from a peer-reviewed 2023 study in the International Review of Law and Economics and a peer-reviewed 2020 event-study in the Journal of Law and Economics, both restricted to Paragraph IV ANDA litigation. Settlement-count data come directly from FTC staff reports on Hatch-Waxman patent settlements for fiscal years 2016 and 2017, the two most detailed counts located for this analysis. Litigation-cost figures come from the AIPLA’s biennial Economic Survey, cited through DrugPatentWatch’s own analysis of that survey’s Hatch-Waxman cost tables. The original taxonomy of file-wrapper exposure types is this article’s own classification, built by generalizing the fact patterns in the two case studies rather than drawn from an existing legal framework, and it should be read as an organizing tool rather than a doctrine courts apply directly. Limitations: settlement-rate statistics from different studies use different sample periods, filters (e.g., first-filer-only versus all filers, active-ingredient patents versus all patent types), and methodologies, and are not directly comparable to one another; they are presented separately rather than averaged together.
Key Takeaways
In Actelion v. Mylan, No. 24-1641 (Fed. Cir. 2026), a stipulated judgment of infringement built on an initial claim construction was vacated on appeal and reversed on remand once the full prosecution history and disclosure-dedication analysis were applied [1].
In Amgen v. Amneal, 945 F.3d 1368 (Fed. Cir. 2020), an unexplained Examiner’s Amendment that narrowed 272 possible binder-disintegrant combinations to 12 was sufficient to bar Amgen’s doctrine-of-equivalents theory against one of three generic defendants, while the other two defendants in the same case reached different outcomes on separate grounds [2].
Brand-firm stakes in Paragraph IV cases that reach a decision average $4.3 billion, against average generic-firm stakes of $204.3 million, an asymmetry that shapes both sides’ incentive to settle [6].
226 to 232 final Hatch-Waxman patent settlements were filed with federal antitrust regulators in FY 2016 and FY 2017, figures that have stayed in a similar range across recent reporting years [7][8].
The number of patents litigated and the years of remaining exclusivity are stronger predictors of settlement than whether the patent covers the active ingredient itself [9].
Median litigation cost for a pharmaceutical patent case with more than $25 million at risk runs around $4 million per side through trial and appeal, a cost structure that pushes both sides toward early settlement estimates before a full prosecution-history review is complete [10][13].
FAQ
What is prosecution history estoppel in simple terms?
It is a rule that stops a patent owner from using the doctrine of equivalents to reclaim claim scope that the owner gave up during patent prosecution, usually by narrowing a claim to get past an examiner’s rejection [3][4].
Does prosecution history estoppel apply to literal infringement or only the doctrine of equivalents?
Only the doctrine of equivalents. Literal infringement is judged against the claim language as issued; prosecution history estoppel only limits how far a patent owner can stretch beyond that literal language [3].
Can a generic company raise prosecution history estoppel before trial, or only at trial?
It is typically raised and argued at claim construction, well before trial, because it turns on the intrinsic record, the claims, specification, and prosecution history, which courts examine at that stage. A strong estoppel argument can also be a factor in settlement negotiations long before either claim construction or trial [1].
What is the disclosure-dedication rule and how is it different from prosecution history estoppel?
The disclosure-dedication rule bars a patent owner from claiming, under the doctrine of equivalents, subject matter that the patent’s own specification describes but never puts into a claim. It does not require any amendment during prosecution at all, which makes it a separate, independent bar from prosecution history estoppel, as the two operated independently in Actelion v. Mylan [1].
How often do Paragraph IV patent cases settle before trial?
Estimates vary by dataset. One academic study of active-ingredient-patent cases litigated to a decision between 1988 and 2012 found settlement rates lower than patent litigation generally, with roughly 45 percent of litigated cases in one related dataset settling rather than reaching judgment. Separately, the FTC recorded 226 to 232 final Hatch-Waxman settlements filed annually in FY 2016 and FY 2017 across all patent types and filer positions [6][7][8].
What is a typical Paragraph IV case worth to a generic filer versus a brand manufacturer?
In cases litigated to a decision, one peer-reviewed event study found average brand-firm stakes of $4.3 billion against average generic-firm stakes of $204.3 million, an asymmetry that shapes settlement incentives on both sides [6].
Can an examiner’s amendment trigger prosecution history estoppel even if the patent owner didn’t originally propose it?
Yes. In Amgen v. Amneal, the narrowing amendment that triggered estoppel was proposed by the patent examiner during an interview and simply accepted by the applicant’s counsel, and the Federal Circuit still applied the presumption of surrender because of what the amendment did to claim scope, not who drafted it [2].
Does an IPR petition create its own kind of estoppel separate from prosecution history estoppel?
Yes, a different one. Under 35 U.S.C. § 315(e), a party that petitions for inter partes review is later estopped in district court from raising invalidity grounds that were raised, or reasonably could have been raised, in that IPR proceeding [11][12]. That is a post-grant procedural estoppel governing invalidity arguments, distinct from prosecution history estoppel, which governs infringement arguments under the doctrine of equivalents and arises from the original examination record.
What should brand manufacturers do during prosecution to reduce future file-wrapper exposure?
Document the actual reason on the record whenever a claim is narrowed, scrutinize Examiner’s Amendments before accepting them rather than treating an interview summary as routine, and be deliberate about which disclosed embodiments and ranges make it into an actual claim rather than being left in the specification alone [1][2].
Where can a generic challenger find a patent’s complete prosecution history before deciding whether to litigate or settle?
The full file wrapper for any issued U.S. patent, including every amendment, rejection, examiner interview summary, and applicant response, is available at no cost through the USPTO’s Patent Application Information Retrieval (PAIR) system. Services such as DrugPatentWatch aggregate prosecution and Orange Book data alongside litigation history for pharmaceutical patents specifically, which can shorten the time it takes to locate the relevant record for a given drug.
References
Crowell & Moring LLP. (2026, May 26). pH, prosecution history estoppel, and patent scope: Three lessons from the Federal Circuit’s latest Hatch-Waxman ruling [Summary of Actelion Pharmaceuticals Ltd. v. Mylan Pharmaceuticals Inc., No. 24-1641 (Fed. Cir. May 13, 2026)]. https://www.crowell.com/en/insights/client-alerts/ph-prosecution-history-estoppel-and-patent-scope-three-lessons-from-the-federal-circuits-latest-hatch-waxman-ruling
Czaplyski, W. (2020, April 1). Federal Circuit clarifies conflicting transitional terms and application of prosecution history estoppel [Case comment on Amgen, Inc. v. Amneal Pharmaceuticals, LLC, 945 F.3d 1368 (Fed. Cir. 2020)]. Harvard Journal of Law & Technology Digest. https://jolt.law.harvard.edu/digest/application-of-prosecution-history-estoppel
Festo Corp. v. Shoketsu Kinzoku Kogyo Kabushiki Co., 535 U.S. 722 (2002).
Warner-Jenkinson Co. v. Hilton Davis Chemical Co., 520 U.S. 17 (1997).
Ajinomoto Co., Inc. v. International Trade Commission, 932 F.3d 1342 (Fed. Cir. 2019).
Jacobo-Rubio, R., Turner, J. L., & Williams, J. W. (2020). The distribution of surplus in the US pharmaceutical industry: Evidence from paragraph IV patent-litigation decisions. Journal of Law and Economics, 63(2), 203-238. https://www.journals.uchicago.edu/doi/full/10.1086/707407
Federal Trade Commission. (2019, May 23). FTC staff issues FY 2016 report on branded drug firms’ patent settlements with generic competitors. https://www.ftc.gov/news-events/news/press-releases/2019/05/ftc-staff-issues-fy-2016-report-branded-drug-firms-patent-settlements-generic-competitors
Federal Trade Commission. (2020, December). FTC staff issues FY 2017 report on branded drug firms’ patent settlements with generic competitors. https://www.ftc.gov/news-events/news/press-releases/2020/12/ftc-staff-issues-fy-2017-report-branded-drug-firms-patent-settlements-generic-competitors
Ahn, K., Trujillo, A., Gibbons, J., Bennett, C. L., & Anderson, G. (2023). Settled: Patent characteristics and litigation outcomes in the pharmaceutical industry. International Review of Law and Economics, 76, Article 106169. https://www.sciencedirect.com/science/article/abs/pii/S0144818823000479
American Intellectual Property Law Association. (2023). Report of the economic survey. https://www.aipla.org/detail/journal-issue/2023-report-of-the-economic-survey
DrugPatentWatch. (2026, March 22). ANDA litigation: The complete playbook for pharmaceutical patent litigators, IP teams, and institutional investors. https://www.drugpatentwatch.com/blog/anda-litigation-strategies-and-tactics-for-pharmaceutical-patent-litigators/
35 U.S.C. § 315(e).
DrugPatentWatch. (2026, April 5). Drug patent litigation costs: The $5M playbook pharma executives actually use. https://www.drugpatentwatch.com/blog/managing-drug-patent-litigation-costs/; DrugPatentWatch. (2025, July 27). How much does a drug patent cost? A comprehensive guide to pharmaceutical patent expenses. https://www.drugpatentwatch.com/blog/how-much-does-a-drug-patent-cost-a-comprehensive-guide-to-pharmaceutical-patent-expenses/
