- with readers working within the Chemicals and Healthcare industries
- within Technology topic(s)
“The reaction chamber therefore was a “black box,” and it was seemingly impossible to prove when, if at all, the patent’s claim elements were satisfied while the accused reduction reaction was in motion. . . ”
I. The Case Inside a Reaction Chamber
Otsuka Pharmaceutical Co. owns JYNARQUE®, the first FDA-approved treatment to slow kidney-function decline in adults at risk of rapidly progressing autosomal dominant polycystic kidney disease. Its active ingredient is tolvaptan. Lupin Ltd. and Lupin Pharmaceuticals, Inc. seek FDA approval to market generic versions of JYNARQUE® tablets through their Abbreviated New Drug Application (“ANDA”) No. 216063. Otsuka sued Lupin in the District of Delaware, asserting patents directed not to tolvaptan as a new compound, but to cleaner ways of making it and to the highly pure tolvaptan those methods produce. See Otsuka Pharm. Co. v. Lupin Ltd., No. 21-cv-00900-RGA, at 10–11 (D. Del. July 31, 2024) (“Bench Trial Op.”).
The appeal in Otsuka Pharmaceutical Co. v. Lupin Ltd., No. 24-2297 (Fed. Cir. May 21, 2026) (“Otsuka, slip op.”) arose from district court Judge Richard G. Andrews's bench-trial judgment that Lupin's ANDA-specified manufacturing process does not infringe the asserted claims of U.S. Patent Nos. 8,501,730 (“the '730 patent”) and 8,273,735 (“the '735 patent”), and that the asserted method claims of the '735 patent are invalid for obviousness. The Federal Circuit affirmed.
Here, we will focus solely on the infringement ruling against Otsuka on the '730 patent—a product-by-process patent claiming highly pure tolvaptan made by a ketone-precursor reduction reaction. How a patentee like Otsuka loses an infringement case it brought is usually a story about claim interpretation or missing evidence. This one is both, and stranger still.
The case turned on the chemical kinetics at work inside a sealed reaction chamber during the minutes before Lupin's process specifications declare the chemical reaction finished. All parties agreed that at the specified finish line, there was no infringement of the '730 patent. To win at trial, Otsuka could not point to the finished process. It had to prove infringement somewhere in the middle of it.
So Otsuka set out to prove its “in-process” theory of infringement—requiring Otsuka to demonstrate what the chemistry is doing at a precise and fleeting moment inside the reaction chamber. To accomplish that feat, Otsuka chose to argue that the plain meaning of the patent claim term “amount” identifies not only a quantity but also a specific moment in time: “practical completion” of the chemical reaction, a concept found nowhere in the patent or its claims.
Let's follow the parties' infringement fight from the claim language to the courtroom. And, in the end, ask whether Otsuka lost the fight because its proof failed or because it set out to prove the wrong thing.
II. The Last Patent Standing
By the time Otsuka faced Lupin's ANDA, the most potent patents and regulatory exclusivities in its portfolio had already expired. The broad compound patent on tolvaptan had expired years earlier, and the method-of-use patent and the orphan-drug exclusivity for JYNARQUE® had also run their course. Bench Trial Op. at 2, 19. Only later-issued and more narrowly focused patents remained.
Otsuka sued Lupin on three of those remaining patents: the '730 patent, the '735 patent, and U.S. Patent No. 10,905,694 (“the '694 patent”). The '694 patent claims certain “amorphous composites” of tolvaptan. Although Otsuka asserted the '694 patent through claim-construction proceedings, it voluntarily dismissed the patent from the case well before trial. The '735 patent survived through the bench trial but was invalidated by the post-trial judgment. This left the '730 patent as the last assertable patent standing in the case.
Unlike Otsuka's original compound patent, which broadly claimed the tolvaptan molecule itself, the '730 patent claims only a tolvaptan molecule that is made by a specific route of synthesis. This means that a competitor like Lupin can make and sell tolvaptan so long as its tolvaptan is made in a way that avoids Otsuka's patented synthesis process.
And therein lies the gist of this case. If Lupin could prevail over Otsuka on the '730 patent, it would be free to sell its generic version of tolvaptan. Otsuka was left with no other remaining asserted patent that could preclude Lupin from going to market.
The '730 patent claims highly pure tolvaptan—99.5% pure tolvaptan (claim 1) and tolvaptan substantially free from at least one of four named impurities (claim 2)—but only when produced by a process involving a specified reduction reaction using a defined amount of hydrogenating agent. See, e.g., the '730 Patent, claims 1 & 2. The chemistry behind the claimed process was straightforward in concept.
The claimed tolvaptan is made by a process that involves reducing a benzazepine ketone precursor to the corresponding alcohol using certain hydrogenating, i.e., “reducing” agents. The reducing agent at issue here was sodium borohydride (“borohydride”).
Reducing ketones to alcohols with borohydride was not new. Reactions reducing ketone precursors in the presence of borohydride have been known since the 1930s. Bench Trial Op. at 17. But the prior-art methods notoriously produced an unwanted dechlorinated impurity as a byproduct of the reaction. See Otsuka, slip op. at 2.
The asserted novelty of the '730 patent was limiting that byproduct impurity by limiting the amount of borohydride used in the reaction. Less borohydride, less impurity. The '730 patent claims translated that insight into numerical ranges: 0.25 to 1.0 molar equivalent of borohydride per mole of precursor in the broader claims, and 0.25 to 0.5 in the narrower claims.
The broader claimed range of borohydride amounts became the entire focus of the dispute. Lupin's ANDA process specified using more borohydride than the claimed range. It specified using “at least” 1.2 molar equivalents of borohydride. See Bench Trial Op. at 11. The '730 patent claimed using “at most” 1.0 molar equivalent of borohydride. That gap explains why Otsuka could not win by merely pointing to Lupin's final process or its final tolvaptan compound.
III. Finding a Way Through a Design-Around
As you can see, Lupin's counter-strategy to the '730 patent was, in part, a carefully selected process specification ensuring the total amount of borohydride used in its process sits beyond the scope of the patent. But read in isolation, that “at least 1.2 molar equivalents” specification fails to tell the whole story. The specification resides in Lupin's Drug Master File (“DMF”) No. 036263, which provides necessary context. See Otsuka, slip op. at 5.
While the DMF describes the entire process flow for Lupin's tolvaptan, “Stage II” of the process is the step relevant to the patent dispute. In Stage II, the “TLV-III” (benzazepine) ketone precursor is first suspended in methanol within a reaction chamber. And although the DMF specification calls for reducing the ketone precursor with “at least 1.2 molar” equivalents of borohydride, the borohydride is not added to the chamber all at once. Instead, the DMF directs the continuous injection of borohydride into the chamber over an extended period, optionally ranging from 30 to 240 minutes. Bench Trial Op. at 11–12; see Otsuka, slip op. at 5.
At the conclusion of the injection period, when at least 1.2 molar equivalents of borohydride have been added, the mixture within the chamber is stirred. Fifteen minutes later, a sample is taken from the chamber and tested for residual ketone precursor. If the residual ketone precursor is 0.05 percent or less, the contents of the reaction chamber are quenched with a mixture of water and hydrochloric acid, and that step of the process is brought to an end. If residual ketone precursor measures above 0.05 percent, then the borohydride injection resumes and a follow-on mixing, sampling, and measuring routine is again conducted seventy-five minutes later. Bench Trial Op. at 12.
Viewed in this broader context, Lupin's adoption of the “at least 1.2 molar equivalents” specification does appear to be a patent-motivated design-around. Yet that design-around motive likely had to strike a pragmatic compromise with the demands of large-scale commercial chemistry.
Presumably, the authors of the DMF process were cognizant of the reaction's dechlorinated-byproduct potential and had to find a way to minimize its occurrence. And although the '730 patent teaches minimizing the amount of reducing agent as one countermeasure, the commercial scale of Lupin's Stage II process may have required further refinements, such as the metered, continuous injection of borohydride over time. If so, Lupin's process may have added a further dimension to the '730 patent's teaching: the unwanted byproduct effects of the borohydride reduction reaction are constrained by incrementing both amount and time.
Whatever the motivation for the continuous borohydride injection specified in the DMF's Stage II process step, one consequence of that process design was an open window of opportunity for Otsuka to prove infringement of the '730 patent. The Stage II process step indisputably finishes its work beyond the patent's scope, but it might very well satisfy all elements of the patent's claims somewhere along the way.
The question was how to prove it. From the case's timeline, it appears that Otsuka may have wrestled with that question until sometime not long before trial.
IV. The “Amount” in Controversy
The problem Otsuka faced was this. Every one of the sampling and testing protocols required by Lupin's DMF is conducted only after 1.2 molar equivalents of borohydride, or more, are already in the reaction chamber. The data collected under the DMF's protocol therefore confirm when the reaction is finished for purposes of running the Stage II process. But they say nothing about where the reduction reaction stood at, for example, a 0.5, 0.7, or 0.9 molar equivalent of borohydride. See Bench Trial Op. at 12.
Given this predicament, Otsuka somehow needed to prove that the tolvaptan produced by Stage II of the DMF process infringed one or more claims of the '730 patent. Claim 1 of the patent, for example, is set forth below:
A highly pure [tolvaptan] having a purity of more than 99.5%, or a salt thereof, which is produced by the process which comprises reducing [benzazepine] in the presence of [borohydride] in an amount of 0.25 to 1 mole per 1 mole of the [benzazepine].
'730 Patent, claim 1 (emphasis added).
The first hint of Otsuka's nascent trial strategy appeared during the claim-construction hearing held on July 1, 2022—roughly one year after Otsuka filed its initial complaint. There, the parties crossed swords over a single claim term in the '730 patent: the word “amount,” as referring to the amount of borohydride reducing agent used in the reaction.
Lupin understandably argued for “amount” to be interpreted as “using a total amount,” a construction aimed conspicuously at the DMF's specified “at least 1.2 molar equivalents,” sitting comfortably outside the claims' scope. See Otsuka Pharm. Co. v. Lupin Ltd., No. 21-cv-00900-RGA, at 4-5 (D. Del. July 26, 2022) (“Markman Op.”). Otsuka, of course, opposed Lupin's proposed construction, asserting that “no construction is necessary” for the claim term and that “amount” should be given its “plain and ordinary meaning.” Id.
In its claim-construction ruling, the district court noted that the parties also disputed how much precision the numerical range carried. Id. Lupin pressed for a strict, exact reading. Otsuka opposed any tightening of the numerical range, perhaps indicating sensitivity to evidence brushing up against the range’s outer limit. The court sided with neither party, rejecting “any construction intended to connote greater precision than ‘1 mole per 1 mole’ or ‘0.25 [moles] . . . per 1 mole’” because the term “suggests neither approximation nor a particular degree of precision.” Id.
Ultimately, the district court agreed with Otsuka that no construction was necessary for the claim term “amount,” and that the term should be understood by its plain meaning. Id. at 5. The court further commented that, “while perhaps not the last word on the subject, [adopting Otsuka's position] does no harm and is more than sufficient at the present time.” Id.
As it turned out, the district court's comment was as prophetic as it was understated.
More than a year later, during the December 2023 trial, the parties engaged in all-out war over the “plain meaning” of “amount.” By this time, Otsuka had decided that “amount” was not only an indicator of a certain quantity but also an indicator of a certain point in time. In its view, “amount,” as used in the claims, could only refer to an amount of reducing agent added while the reduction reaction was kinetically active. By Otsuka's logic, any amounts added after completion of the reaction would play no role in the reaction itself and would simply be superfluous.
While not receptive to the notion of welding a temporal limitation onto the claim term “amount,” Lupin argued that, by Otsuka's logic, “amount” should mean the quantity of reducing agent added up to “absolute completion” of the reduction reaction.
With an apparent strategy now in mind, Otsuka responded that, in the real world, Lupin's asserted “absolute completion” of a chemical reaction is merely a theoretical possibility. According to Otsuka, the proper temporal gauge by which to measure “amount” is the more nuanced moment of “practical completion” of the reduction reaction.
While careful to note that it was not revisiting its claim construction, the district court was ultimately convinced that “absolute completion” is achievable only in theory. The kinetics of a chemical reaction tend to slow, and even grind to a halt, as the reaction's raw material is consumed. The court therefore agreed with Otsuka's apparently more pragmatic assertion of “practical completion.” Bench Trial Op. at 13–14.
After the battle over the plain meaning of “amount” was decided in Otsuka's favor, the question was no longer how much borohydride Lupin ultimately poured into the reaction chamber—the 1.2-plus equivalents that clearly sat outside the scope of the claims—but how much had been added by the time the reaction was “practically complete.” In winning this specific battle, Otsuka both neutralized the DMF's express specification of “at least 1.2 equivalents” as controlling the infringement inquiry and secured a pathway to presenting evidence of what transpires kinetically inside the reaction chamber during the contested reduction reaction.
Otsuka fought for, and won, the right to present evidence beyond the express product specifications in Lupin's ANDA and DMF. The only question remaining was whether this hard-won theory of proving infringement could carry Otsuka to victory.
In the end, securing the right to present a theory of in-process infringement was one thing. Proving it turned out to be something else entirely.
V. A “Practically Incomplete” Trial
A. The Black Box
In addition to the specific process limitations of the '730 patent claims, Otsuka had to prove that Lupin's process produced either (a) tolvaptan “having a purity of more than 99.5%” (independent claim 1) or (b) tolvaptan “which is substantially free from at least one of” four identified impurities, Compounds A through D (independent claim 2). Here, Lupin extended a superficial olive branch. It conceded that the tolvaptan produced and specified by its DMF process met both claims' purity limitations. See Undisputed Facts ¶¶ 41, 48; Tr. 64:19–68:16 (Roush). But ultimately that concession did little to ease Otsuka's burden of proving infringement.
Otsuka faced two problems. First, it had Lupin's concession that the DMF process produced the highly pure tolvaptan specified in the '730 patent's claims. But the DMF protocol insured that the tolvaptan extracted from the reaction chamber and certified for purity was, in some way, involved in a reduction reaction in the presence of at least 1.2 molar equivalents of borohydride—beyond the scope of the claims' process limitations.
Second, the DMF protocol insured that nothing relevant to Otsuka's evidentiary needs was extracted from the reaction chamber and tested until after at least 1.2 equivalents of borohydride had been added. The reaction chamber therefore was a “black box,” and it was seemingly impossible to prove when, if at all, the patent’s claim elements were satisfied while the accused reduction reaction was in motion.
This explains why Otsuka fought so hard to hitch “practical completion” of the reduction reaction to the plain meaning of “amount” in the patent's claims. In other words, if Otsuka could somehow prove the reduction reaction reaches practical completion before more than 1.0 molar equivalent of borohydride has been added to the reaction chamber, it just might prove Lupin’s infringement of the patent.
So the case boiled down to a single issue: whether Lupin's ketone-precursor reduction reaction reaches “practical completion” in the presence of a claimed “amount” of borohydride—and no more. Because the events comprising this issue transpire entirely within the reaction chamber's “black box,” Otsuka turned to its testifying expert, Dr. William Roush, to sketch into this evidentiary blind spot his version of those events.
B. The Expert's Touch
Dr. Roush offered the court a colorful description of the reaction-chamber chemistry. His account conjures images of an old seafaring voyage: well provisioned and boisterous at the outset but fading into quiet starvation at its end.
He described the chemistry as molecules that must “bump into one another”; plentiful and strident in the beginning, the free-ranging ketone-precursor molecules eventually dwindle in number to a “needle in a haystack” that would take “years and years and years”—“geologic ages” even—to find and “bump into” any added borohydride molecules. Tr. 102:10–103:17 (Roush). A reaction is “complete in a practical sense,” he testified, long before that molecular destitution occurs. “Absolute completion” is “only a theoretical possibility,” he opined, boldly adding, “No organic chemist has ever had a reaction go to absolute completion.” Id.; Tr. 147:10–148:1 (Roush).
Dr. Roush augmented his opinion testimony with data from two real-world experiments conducted by Lupin while developing its DMF process—Experiments 109 and 115. These experiments provided the only data in the record tracking the ketone precursor's disappearance over the course of the reduction reaction. Tr. 92:1–11, 145:14-25 (Roush). In each, after 0.91 molar equivalents of borohydride had been added, just 0.01 to 0.02 percent of the precursor remained, and adding a further 0.39 molar equivalents (to a total of 1.3 molar equivalents) changed nothing. Tr. 80:13–88:24 (Roush).
The “0.91 molar equivalents” in this evidence helps explain Otsuka's resistance, at claim construction, to any hardening of the amount ranges. In laboratory experiments with real margins of measurement error, 0.91 molar equivalents sit uncomfortably close to the claims' outer limit of 1.0 molar equivalent.
Undaunted by this evidentiary “close call,” Dr. Roush presented those data as proof that the reduction reaction was, for all practical purposes, practically completed within the claimed range of borohydride amounts. Tr. 74:18–75:7 (Roush). And perhaps either to widen the uncomfortable gap between 0.91 and 1.0 molar equivalents or to hedge against conditions in the two experiments that did not match conditions in the DMF reaction chamber, Dr. Roush extended his opinions about the experimental data even further.
Whatever the motivation, it was an artful move intended to translate apparent weakness into advantage.
He noted that Lupin's DMF process uses both higher input concentrations and higher temperatures than were used in Experiments 109 and 115. According to Dr. Roush's computations, the DMF process would accelerate the reduction reaction roughly four times faster than the more pedestrian experimental reactions.
This meant that the ketone precursor in the DMF process would be consumed in a practical sense “at least by the time 0.5 molar equivalents” of borohydride were added. Tr. 93:25–95:20, 108:7–109:23 (Roush). And 0.5 molar equivalents sit much more firmly within the claims' 0.25 to 1.0 molar-equivalent range than the experiments' own 0.91 molar equivalents.
With the experimental data recast in this new light, Otsuka was set to argue in post-trial briefing that the reduction reaction in the reaction chamber is practically complete well before 1.0 molar equivalent of borohydride—the claims' outer boundary—is ever added.
This was an elegant rendition of Otsuka’s case theory married to an appealing rationalization of the evidence. One can understand Otsuka’s devotion to it. The only remaining question was how this would play with the district court?
C. Proving Too Much?
As tidy as Dr. Roush's computations appear, the district court bought none of it.
Before getting to the details of the district court's decision and its specific findings, let's venture out on our own and consider a little more carefully what Otsuka effectively proved at trial. Its efforts may well have been self-defeating.
It is true that in Experiments 109 and 115, only 0.01 and 0.02 percent of the ketone precursor remained after adding 0.91 molar equivalents of borohydride, and those percentages remained unchanged even after adding up to 0.39 more molar equivalents. This might very well infer a “practical completion” of those experimental reactions at 0.01 and 0.02 percent residual ketone precursor.
But recall that Lupin's DMF specifies terminating the reduction reaction when residual ketone precursor reaches 0.05 percent. Whether inadvertent or not, Otsuka succeeded in proving the reduction reaction can proceed to as low as 0.02 and even 0.01 percent residual ketone precursor under ordinary laboratory conditions. Otsuka may have over-shot the target.
The burden Otsuka assumed in winning its novel claim construction was proving that the DMF's reduction reaction would reach “practical completion” mid-process. Instead, it may have proven that the reduction reaction in the DMF process is terminated before the reaction is “practically complete” (even after at least 1.2 molar equivalents of borohydride have been added). According to its own trial evidence, Otsuka proved that the reduction reaction standing at the DMF's specified 0.05 percent residual ketone precursor has the potential to further reduce as much as 0.04 percent more of the residual precursor, down to only 0.01 percent.
D. Proving Too Little
Consistent with these arm-chair observations, the district court bluntly found that “Otsuka . . . fails to prove its assertion that the reaction reaches practical completion when 0.05 percent of the ketone precursor remains.” Bench Trial Op. at 14. This finding was fatal to Otsuka's case. More to the point, it was fatal to Otsuka's framing of the case.
The district court apparently viewed the DMF's 0.05 percent residual-ketone-precursor specification as an evidentiary benchmark for an obvious reason. It was the only measurement in evidence—concerning not the laboratory experiments, but the accused DMF process itself—that could be used to determine the kinetic state of the reduction reaction at a particular moment in the reaction's life.
If, after the addition of at least 1.2 molar equivalents of borohydride, the reaction is still active at 0.05 percent precursor residuals, then Otsuka obviously loses. In contrast, if the reaction is proven “practically complete” and kinetically stalled at 0.05 percent precursor residuals, then Otsuka should be heard further on its evidence that Stage II of Lupin's process achieves 0.05 percent precursor residuals before more than 1.0 molar equivalent of borohydride has been added.
Given these stakes, one would think Otsuka presented substantial evidence to prove that the DMF's specified measurement of 0.05 percent residual ketone precursor was correspondingly a measurement of the reaction's “practical completion” under its own theory of the case. Apparently, it did not.
The district court found that Otsuka “fail[ed] to prove its assertion” because Otsuka failed to present any real-world evidence that the reaction is practically complete at 0.05 percent residual ketone precursor. Instead, Otsuka had attempted to shift the court's attention to an alleged party admission. Otsuka argued that the DMF and “Lupin chemists consider the reduction reaction complete” at 0.05 percent residual ketone precursor. Otsuka Proposed Findings at 5-6 ¶ 23 (emphasis added).
The court rejected this asserted party admission as a stand-in for actual proof of “practical completion” at 0.05 percent. In doing so, the district court concluded that the DMF says nothing about the kinetic completion of the reaction, finding instead that the DMF's “0.05 percent operates as an acceptance criterion . . . , not the point at which chemists no longer observe any further reduction.” Bench Trial Op. at 14.
There was logic to that finding. Lupin's expert, Dr. William Dichtel, explained that the 0.05 percent acceptance criterion was motivated by a pragmatic decision to “move on so as to expedite the synthesis of [tolvaptan]” “and tolerate whatever leftover [ketone precursor] there might be [because] the reaction has gone to a point where it's in [Lupin's] economic interest to move on to the next step.” Tr. 286:20–291:4 (Dichtel).
In other words, the 0.05 percent acceptance criterion was grounded more in the laws of economics than in the laws of chemistry. Although there might yet be enough molecules left in the reactor to “bump into each other” and further the reaction, it is not worth commercial production time to let the molecules finish their work—Dr. Roush testified, after all, that finishing the reaction's work can take a long time, “geologic ages” even.
And we shouldn't easily discount the likelihood of sound technical reasons for terminating the reduction reaction at the chosen 0.05 percent endpoint. Recall that the asserted novelty of Otsuka's patent is decreased reaction-byproduct impurities, achieved by limiting the amount of borohydride used in the reaction. Terminating the reaction at the kinetically incomplete 0.05 percent acceptance endpoint might well achieve a similar benefit—preventing the production of additional byproduct impurities by ending the reaction altogether.
E. A Belt and Suspenders
The district court buttressed its judgment of no infringement with an alternative finding: “Even assuming Lupin's reduction reaction reaches practical completion at the 0.05 percent ketone precursor level, Otsuka fails to show that no more than 1 molar equivalent of [borohydride] or less has been added by that point.” Bench Trial Op. at 14.
The district court noted Otsuka's reliance on Experiments 109 and 115 but was unmoved by that experimental evidence because it was conducted under different conditions than the DMF process. The court acknowledged Dr. Roush's attempt to equate the experiments with the DMF process through calculations showing that “[b]ased on the differences in starting ketone precursor amount and temperature, . . . Lupin's DMF process would proceed 4-fold faster than Lupin's experiments.” Id. at 15 (citing Tr. at 93:16–95:20, 98:16–99:25, 109:7–23).
But the court rejected Dr. Roush's opinion that other differences, including scale, pH, and solvent conditions, “were immaterial.” Id. (citing Tr. at 91:23–98:13). The court instead credited Dr. Dichtel's testimony that “these differences were meaningful and that given the closeness between 0.91 molar equivalents and 1 molar equivalent and the decreased speed of the reaction as it nears completion, even small discrepancies in Dr. Roush's assumptions could cause this practical completion point to fall outside the claimed range.” Id. (citing Tr. at 197:6–24).
Most devastating to Otsuka's infringement case was evidence that the experimental data were themselves unreliable. On cross-examination, Dr. Roush conceded an unexplainable anomaly in Experiment 115. After the full charge of borohydride was introduced into the reaction, the previously declining ketone precursor appeared to regenerate and suddenly increase in amount—a chemical impossibility. Id. (citing Tr. at 88:9–16, 89:14–25, 118:3–120:12, 121:1–22, 123:10–23); see also Otsuka, slip op. at 9.
Dr. Roush opined that these “anomalies . . . might be accounted for by some unspecified ‘margin of error’ or a lack of appropriate quality control.” Bench Trial Op. at 15 (citing Tr. at 88:9–16, 89:14–25, 118:3–120:12, 121:1–22, 123:10–23). The district court rejected those attempted justifications, finding instead that Otsuka failed to prove “by a preponderance of the evidence that Experiments 109 and 115 provide reliable results that are representative of Lupin's DMF process.” Id. (emphasis added).
F. Begin with the Endpoint in Mind
In view of Otsuka's reliance on the two anomalous experiments, the district court made its single most glaring critique of Otsuka's case: “Otsuka conducted no experiments to replicate Lupin's process.” Id. at 15 (citing Tr. at 100:2–14, 124:2–4) (emphasis added).
Was this a forfeited evidentiary opportunity? If Otsuka had conducted small-scale replications of the DMF process that produced favorable data, the district court just signaled its willingness to have received them in evidence at trial. Presumably such data would present none of the reliability issues that plagued Experiments 109 and 115. And the outcome of the case might have been different.
Whether Otsuka's consulting experts ran such experiments under the auspices of attorney work-product immunity will never be known. If they did, we can assume the results were not positive for Otsuka.
But we have to wonder: if any such experiments were conducted, were they designed with the right endpoint in mind?
Otsuka pinned its fate to proving “practical completion” of the patent's reduction reaction—a notion found nowhere in the claims or the patent itself. The one thing Experiments 109 and 115 made clear is that the reduction reaction can be efficiently driven as low as 0.02 and 0.01 percent residual ketone precursor—a wide miss of the DMF's specified 0.05 percent termination point.
But nothing in the patent's claims required Otsuka to prove the accused reaction is brought to “practical completion.” That was an endpoint burden of proof Otsuka assumed all to itself. For its part, Lupin does not even hold itself to the burden of pushing the Stage II process to practical completion. Why should Otsuka do so in proving its case?
Perhaps Otsuka would have been better served by tying both its evidence and its arguments to a different endpoint event. There was one such endpoint sitting quietly in the case all along.
Dr. Dichtel testified that Lupin's commercial objective is economically producing pure tolvaptan, not driving the reduction reaction to practical completion. And producing pure tolvaptan also happens to be an objective shared by the '730 patent—expressed as 99.5% purity in claim 1 and “substantial freedom” from four known contaminants in claim 2. It would seem that production of pure tolvaptan, as defined by the claims themselves, is the right experimental endpoint.
And keep in mind, the '730 patent claims “producing” “highly pure tolvaptan” by a “process.” Shouldn’t the claims’ endpoint correspond to completion of that process?
If so, it follows that any quantity of “highly pure tolvaptan” “produced” in the presence of a claimed amount of borohydride marks completion of the claimed process. Said differently, the endpoint of the claimed “process” is the production of its intended product—highly pure tolvaptan.
The reduction reaction recited in the claims is a means to that end, not the end itself. After all, the point of the minimized amount of borohydride specified in the claims was minimizing byproduct impurities, not driving the reaction to practical completion.
Does Lupin's Stage II process produce pure tolvaptan, as defined by the claims, sometime before more than 1.0 molar equivalent of borohydride has been added—regardless of whether the reduction reaction is “practically complete”? None of us know for sure.
Perhaps that precise experiment has never been done. But if it were, the experiment would be designed with a focus on the patent claims themselves and with the right endpoint in view. And it would more faithfully resolve the infringement inquiry.
Maybe someday a skilled chemist with enough curiosity and a little free time will do the work and find the answer, one way or the other. Until then, one can't help but feel that the trial here was somehow left “practically incomplete.”
In the meantime, the district court found that “Lupin's DMF process does not infringe the asserted claims of the '735 and '730 patents,” due to Otsuka's failure “to prove by a preponderance of the evidence that the reduction reaction in Lupin's tolvaptan synthesis process reaches completion before more than 1 molar equivalent of [borohydride] has been added.” Id. at 15–16 (emphasis added).
The Federal Circuit affirmed the district court's findings under clear-error review. Otsuka, slip op. at 8–10.
VI. The Winning Needle
Dr. Roush’s vivid description of the reaction chamber’s inner workings might evoke familiar imagery for attorneys who litigate patent cases. At the outset, as discovery unfolds, myriad theories of the case emerge and are tested, much in the same way as Dr. Roush’s kinetically raucous molecules “bump into each other,” testing for fit and compatibility. But unlike Dr. Roush’s reactions, lawsuits march toward a moment of absolute completion. And while claim construction and the intervention of experts might help in selecting one theory over another, only one winning needle can be plucked from a haystack of imposters. Invariably, that winning needle is the one that faithfully threads together each and every limitation of the patent’s claim.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
[View Source]