Obviousness and Stereochemistry

Amgen Inc., v. Sandoz Inc. 22-1147, — F.4th — (Fed. Cir. Apr. 19, 2023)

Jordan is a second-year law student at the University of Missouri and a registered patent agent.  He has an extensive background in chemistry and food science.

Amgen markets apremilast, a phosphodiesterase-4 (“PDE4”) inhibitor, which is used for treating psoriasis and related conditions, under the brand name Otezla® which is covered by three patents, U.S. Patents 7,427,638, 7,893,101, and 10,092,541. Sandoz submitted an Abbreviated New Drug Application (“ANDA”) seeking approval market a generic version of apremilast. Celgene, the original plaintiff, brought this Hatch-Waxman suit, asserting that Sandoz’s generic product would infringe the ’638 and ’101 patents. The Federal Circuit affirms the district court’s findings on all issues raised.

Sandoz asserts that U.S. Patent 6,020,358 renders the Amgen patents obvious. The ’358 patent is the first U.S. patent describing a racemic mixture containing apremilast. Enantiomers detail the orientation of the molecule around a chiral center and can have vastly different therapeutic outcomes and clinical results between the (+) and (-) enantiomers, particularly thalidomide analogues. The critical difference between the ‘358 and ‘638 patents is the composition of the mixture:

The ’358 patent is a racemic mixture comprised of 50% of the (+) enantiomer of 2-[1-(3-ethoxy-4-methoxyphenyl)-2-methylsulfonylethyl]-4acetylaminoisoindoline-1,3-dione and 50% of the (-) enantiomer.

Compare to an excerpt of Amgen’s language in claim three of ‘638:

A pharmaceutical composition comprising stereomerically pure (+)2-[1-(3- ethoxy-4-methoxyphenyl)-2- methylsulfonylethyl]-4- acetylaminoisoindoline-1,3-dione, or a pharmaceutically acceptable salt, …

The ’638 Patent: Sandoz failed to provide clear and convincing evidence at the district court to show that a skilled artisan would have reasonably expected a benefit from separating the enantiomers or that the (+) enantiomer was the cause of the desirable properties of the ‘358 formulation. Sandoz’s appeal asserts that the district court erred in its findings.

At the Federal Circuit, Judge Lourie, writing for a unanimous panel of Judges Cunningham and Stark, focuses on the unexpected potency of apremilast discovered relative to the apremilast-containing racemic mixture during testing and experimentation. At the district court, the named inventor, Dr. Schafer, testified that the stereomerically pure apremilast reduces the production of tumor necrosis factor alpha (“TNFα”), the factor that is linked to psoriasis, 20 times more effectively than the previous patent.

Dr. Schafer also says that a skilled artisan would expect a twofold improvement in efficiency. The district court finds that the 20-fold difference, when an otherwise two-fold difference would have been expected by the skilled artisan was sufficient to support a finding of an unexpected result and subsequent nonobviousness. A 20-fold difference from the steromerically isolated formulation goes well beyond a difference in degree into a difference in kind.

It strikes me as interesting that the characteristics of apremilast were largely left undiscussed. Apremilast is a thalidomide analogue and safety concerns about the infamously teratogenic effects related to thalidomide were sufficient skepticism for a skilled artisan and the industry at large regarding stereomerically pure formulations. However, it would seem that this teratogenicity could cut both ways as a skilled artisan would have ample incentives to investigate the enantiomers separately for safety concerns in light of the thalidomide paradox.

The thalidomide paradox provides that racemic mixtures and pure enantiomers have different characteristics, despite in vivo racemization. If the molecule is sufficiently analogous to thalidomide, it seems that a skilled artisan would be inclined to be taught towards separating the enantiomers. Regardless, the unexpected result was found to be dispositive on appeal and the objective indicia weren’t necessary but the other findings of the objective indicia of long-felt need, others had tried and failed, industry skepticism, and commercial success were also affirmed.

           The ’101 Patent: Sandoz also alleges that the ’101 patent should have been found to be obvious. Sandoz says the ‘101 patent should not be entitled to the priority date of March 2002 because the  ’515 provisional application did not inherently disclose crystalline Form B of apremilast and therefore did not satisfy the written description requirements of 35 U.S.C. § 112. The district court relies on a finding that ‘515 application did inherently disclose the crystalline form.

The Federal Circuit did not find it necessary to delve into inherent disclosure because a review of the trial record shows Amgen provided multiple experiments, and expert testimony, showing that there was actual disclosure of the crystalline Form B. Absent any contrary evidence by Sandoz, the priority date and subsequent finding of nonobviousness were affirmed. Another win chalked up to Amgen’s excellent expert testimony.

            The ’541 Patent: On cross-appeal, Amgen asserts that the district court finding that the patent is obvious in light of prior art is erroneous. The district court found that the ’541 patent would have been obvious over the prior art or in crediting expert testimony stating that dose-titration modification would have been routine to a skilled artisan. Specifically, dose titration would be something that is regularly done in the treatment of psoriasis.

32 thoughts on “Obviousness and Stereochemistry

  1. 3

    Strange case. Case law on the obviousness of an enantiomer in view of its stereoisomer goes way back. It’s obvious to try to purify it and IMHO overcoming the presumption that the pure enantiomer is obvious would require a showing that purification was (for whatever reason) deemed to be nearly impossible or that the pure enantiomer would be poison. 2 fold or 20 fold more “activity” in some test is irrelevant to the obviousness of the compound itself if the purification is obvious to try.

  2. 2

    I don’t understand why 20x versus 2x really matters as an unexpected result. The ordinary artisan would already have expected, according to testimonial evidence from the named inventor, a 2x increase in effectiveness. Is that, by itself, not reason enough for the ordinary artisan to try the well-known process of separating the enantiomers?

    The inquiry for unexpected results should be whether the predictable result was itself a reason for the ordinary artisan to try what the prior art has taught. If they get an unexpected result beyond that, that’s great, but the ordinary artisan was already taught to make the invention and would have inevitably found the unexpected result anyway.

    1. 2.1

      The inquiry for unexpected results should be whether the predictable result was itself a reason for the ordinary artisan to try what the prior art has taught. If they get an unexpected result beyond that, that’s great, but the ordinary artisan was already taught to make the invention and would have inevitably found the unexpected result anyway.

      I think abstracting the legal principle so far out from specifics is not terribly helpful. I suppose that there might be situations in which your maxim is right, but I can easily think of others where that maxim does not usefully advance the ends that patent law is meant to serve.

      It is well known, for example, among oncologists that if you are going to prescribe a two drug treatment regimen, then it is best to use two drugs that act through different molecular channels, because the results will often be more than additive (aka synergistic). Imagine, then, that platinum drugs (in general) and nucleoside analogs (in general) achieve (in general) a degree of control that is twice what you would expect from adding their individual efficacies.

      Further imagine that it turns out that the particular pairing of carboplatin (a platinum drug) + AZT (a nucleoside analog) achieves a 20-fold better degree of control than would be expected from summing their individual reported effects.

      It could take the rest of the oncology field a lot of time to discover that unique advantage one-by-one until it became a well known fact in the field. You really should want the first oncologist who discovers this felicitous fact to disclose it to the rest of the field. Patent law is the legal mechanism by which we encourage this disclosure. If we set the standard for obviousness such that such disclosures are never patentable, then at the margin we slow the spread of such adventitious discoveries.

      1. 2.1.1

        The consensus academic theory is that the obviousness bar should be set at a point such that protection is only offered to advances which would not have occurred without the patent incentive.

        You could make a theoretical argument that this bar should presumptively be low for pharmaceutical inventions, given unique market dynamics: 1) regulatory entry barriers that raise fixed costs of commercializing inventions, even small improvements; 2) tough mandatory substitution laws that force intense price competition between brands and generics; 3) high costs of experimentation to determine whether improvements are useful or not.

        In a nutshell, no one is going to do anything new unless they have some exclusivity attached, even if it’s a minor improvement. Maybe they would pursue new indications regardless, but categorically denying patentability to new MOT patents for these new applications would actually reduce the drug’s effective exclusivity (if other MOT patents are valid), and would be horrible policy.


          That’s a slippery slope you are on there, and while “enough to recoup” is certainly one of the rationales, it is by NO means the only one.

          (dont fall for the “must make” type of propaganda — and don’t forget that the word “promote” in its legal context includes the (mere) advertise, as in ‘advertising promotion.’


          Kyle, I read your post with astonishment. Do you have a citation to back up your assertion about a consensus in academia? What makes you assert that “no one” is going to do “anything” new unless the State provides “exclusivity”. I doubt it very much.


            I think that he means no one in the pharma industry is going to do anything new without exclusivity.


              Another hidden link (“t”), has the propagandist Drum “blaming Republicans”…..


              Perhaps it is the sheer inanity that prompts Greg to hide these hyperlinks.


              If you mean that “he” means academia, then MaxDrei’s call is still in order.

              Knowing the bias of Pharma may alleviate the audacity (as well as recognizing that “academia” lacks any ethical restraint between teaching the law as it is and how they (the Royal “They”) would want the law to be).


            1) Abramowicz and Duffy, the Inducement Standard of Patentability (2011). But it’s kind of just intuitive—why offer a patent for an inevitable development? Why worry about giving a patent to a development that would not have existed otherwise? The price of something never developed is infinity.
            2) I was referring to pharma specifically. I can’t recall a single improvement to an approved molecule without patent protection, other than some new indications (referred to above)


              Kyle: “I can’t recall a single improvement to an approved molecule without patent protection, other than some new indications (referred to above)”

              This seems like something that could be checked and verified instead of relying on one’s memory. I would be shocked if your memory accorded with reality here.



                You assail Kyle for something that (he) could easily do, then turn around and…

                do the exact same thing



                The typical extended release or “dissolving” dosage form is what I had in mind here. Often simple application of familiar principles to reformulate a known molecule, but if it could be genericized immediately, I doubt it would pencil out for the brand.

                As a matter of patent theory, I don’t know what the right answer is. Seems like KSR, and its centuries of predecessors, don’t give any slack to inventions in peculiar markets. But maybe the law should?

                1. But maybe the law should?

                  Ever hear of ‘divide and conquer?’

                  Dividing out the one section that naturally wants strong patent protection is possibly the worst possible move to combat Efficient Infringers and would damage destructive innovators.

                  No thank you.


                Exclusivity for small molecule (not biologics) brand drugs without patent protection is 5 years. That means, generics can come on the market in 5 years. Patents offer much longer exclusivity.

                Exclusivity for large molecule brand drugs is currently 12 years. A patent may or may not offer longer exclusivity, but there is also patent term extension available for up to 5 years, if there are fewer than 14 years of patent exclusivity left.

                It is logical.

      2. 2.1.2

        You make a good point. I should add a caveat to my earlier statement that relates to the specific case at hand, namely, that there are only two enantiomers, rather than some large and possibly open-ended collection of things to choose from. This gets back into the KSR “obvious to try” rationale, which is usable when you have a finite (and typically small) number of identified options to choose from, and there is some reason established in the prior art to try them in some form.

        In the apremilast example, with necessarily just two enantiomers, along with an acknowledged 2x benefit of separating them and using whichever one is actually bioactive, the ordinary artisan would presumably be inclined to do that already. The 20x, rather than 2x, actual benefit arises even though the ordinary artisan just did what he was taught to do by the prior art.

        In your carboplatin + AZT example, a claim to a generic platinum drug + nucleoside analog combination would not be patentable because the 2x benefit is already expected. But the specific combination that has an unexpected result is discovered not through a well-known step already taught by the prior art, but rather by an extensive process of trial and failure.

        In terms of the reward given and the incentive provided to inventors, I think that really gets at the crux of the current case: does it incentivize innovation to tell the general public, hey, you don’t get to do the well-known process of enantiomer separation on this well-known drug because it turns out to work better than expected? I personally don’t think it does, because (absent any other evidence to the contrary) it was a likely thing that someone was going to try sooner or later, and the only reason they might not try it is because the racemic mixture was already subject to a patent.

        Meanwhile, I agree with you that the discovery of a particular combination of drugs drawn from a large and possibly open-ended collection of choices, as in your example, should be incentivized, to entice inventors into exploring those possibilities when they otherwise might not.


          In reality, both this patent and the prior art patent should be invalid for lack of enablement or utility. These lab bench tests are rarely if ever predictive of efficacy, and their disclosure does not enable a POSA to administer the compound in a dosage form. As a matter of real-world innovation maximization, allowing the patentee to tie up these compounds disincentivizes others from investigating uses for them and could actually delay any of them from coming to market as drugs.


            In reality, both this patent and the prior art patent should be invalid for lack of enablement or utility.

            In reality, there is an exceptional number of Pharma patents that should be invalid for lack of enablement AT the time of filing (one only has to look at the number of items already filed that
            FDA testing for their claimed utility.

            But you won’t hear a word of this from Greg “I-Use-My-Real-Name-Except-When-I-Don’t” DeLassus.


              Where in Section 101 is the utility tied to FDA approval? And in this case, if we agreed to your premise, the titation patent (the 541) in the case, then would not be invalid, because a prior art reference must be enabling, and if the titration regime wasn’t tested, it can’t be obvious, right?


                Utility is tied to the “as claimed”

                (This is a slight modification to my prior cries on this account)

                1. SO in this case, you have an enantiomer of a PD4 inhibitor. Where’s the beef with enablement?

                2. My comment is not necessarily about this case — as should be clear in context of my comment’s “exceptional number.”


                It seems to me that in reality, pharma claims fall broadly into two buckets: (1) composition claims; and (2) method of treatment claims.

                If a composition is disclosed to (e.g.) inhibit insulin growth factor, but does so only weakly (such that the FDA will not approve it for human use), this does not really negate its enablement or utility in the patent law sense. The claim just says “a compound according to formula…,” with no promise about an activity. The utility is disclosed in the specification, not in the claim, and so long as the compound really does have such a utility (even if only slightly), then the utility is satisfied.

                When it comes to methods of treatment, unless the claims explicitly limit the target patient (e.g., “a method of treating fleas in a canine” or “a method of treating diabetes in an adult human”), then the proper reading of the claim scope should be to read on the claimed method being useful in human individuals without regard to age, sex, etc. In such a case, there are two possible outcomes here.

                In the first, the claimed method is not effective enough for the FDA to approve it. I suppose, then, that it would be fair to say that the claim suffers a scope of enablement defect. This tends not to matter much in practice, however. If the FDA will not approve the method, then it will not be infringed in any detectable fashion, in which case the invalidity of the claim will never matter, because it will never be asserted.

                If the second scenario, the method does work well enough for the FDA to approve it. In that case, where is the supposed enablement or utility defect?

                In other words, I just do not see a problem here that some claims cover therapies that fail FDA review. Some of those claims are valid, others are not, but none of them will be asserted in any event, so their validity or lack thereof will never be a matter of more than purely academic interest.

                1. In scenario 1, it may block others from pursuing the same or a similar drug/indication. In scenario 2, it may enable evergreening—though this is less often a concern than people imagine.

                  Also, the FDA will approve a drug if it has *any* efficacy. It need not be more effective than alternatives; your hypo assumes the conclusion by positing a spec that discloses certain but weak efficacy. In reality, typical compound patents disclose info that very very slightly enhances the likelihood of efficacy from like 0% to 1%.


                Brenner said you must disclose “substantial utility” “in currently available form”. Robinson S198 has similar language: “capable of immediate useful application”.

                In my view, The 541 may still be invalid over prior art Phase 2 and 3 clinical studies establishing safety and efficacy in treating psoriasis.

                1. method appears to be a filing because of

                  I think that THAT was his point — that later filing is obvious due to the published FDA information.

                  Yes, one’s own work can render obvious one’s later filings.

                  Sure, I “get” this may well be a systemic problem that certain items are forced through the FDA, but ALL areas of innovation require possession at time of filing, and — literally — if one is awaiting testing results then one does not posses to HAVE an earlier filing date.

    2. 2.2

      Because the inventor still “discovered” something that was not obvious (the discovery that the combination worked 20x better). Although agreed that it doesnt fit that easily into 103 language or jurisprudence…

      1. 2.3.1

        Of course you are.

        (easy check: does the view go against the patent holder?)

        But nooooooo, Malcolm is not anti-patent (as the gaslighting goes on…)

  3. 1

    Slightly off topic, but using the transitional phrase “comprising” when a key part of the claim is that a certain component be “pure” is somewhat problematic. Purity means that other things are not present. Comprising means other things can be present.

    It reduces to the same problem as the razor case, which was something like “A razor comprising a blade assembly, the blade assembly consisting of no more than 3 blades” and the court found a 4-blade razor read on the claims, because of the “comprising” limitation.

    A racemic mixture “comprises” crystals of stereomerically pure (+) compound. It also comprises crystals of stereomerically pure (-) compound. It comprises both. Its somewhat unsetting language, but not sure how you get around it.

    1. 1.1

      Not sure I agree, as “pure” may be open to non-active items also present.

      In other words, “pure” may only be related to other active things are not present.

      I did not recall off the top of my head that razor case – do you mean this:
      link to casetext.com

      Of which, this line may be differentiating: “The inventive contributions of the ‘777 patent are varying progressively the exposure and spacing parameters of the blades to overcome the undesired drag forces produced by razors with multiple blades, not simply limiting the number of blades to three.

Comments are closed.