Infringement under the Doctrine of Equivalents: There is life in Q3

Reading time: 13 minutes
DISCLOSURE NOTICE: Hepp Wenger Ryffel is involved in this matter on behalf of Mepha Pharma AG.

Case Nos. O2021_004, O2021_005 | Judgment of 20 April 2023 | ‘Deferasirox II’

HEADNOTES

Art. 229 para. 1 CPC
Proper nova
The decision of a Board of Appeal of the European Patent Office which also concerns the patent in suit is a genuinely novel fact within the meaning of Art. 229 (1) (a) of the Code of Civil Procedure.

Art. 8 CC, Art. 1 para. 2 PatA, Art. 56 EPC
Burden of proof
The burden of proving the technical effect(s) of the invention lies with the patent proprietor.

Art. 51 PatA, Art. 66 let. a PatA, Art. 69 EPC
Infringement by equivalent means
If a preferred value is mentioned in the patent and a range is claimed in the claim that includes this value, the skilled person assumes that the patent proprietor has waived protection for values outside the claimed range.

This judgment in litigation between Mepha and Novartis covers both Mepha’s revocation action with respect to EP 202 and EP 018 (O2021_004) and Novartis’ conterclaim for infringement of both patents (O2021_005) by Deferasirox-Mepha® (Swiss market authorization no. 67678).

Fe(III) chelate

Deferasirox is an iron chelating agent; it is frequently referred to simply as DFX. It is mainly used to reduce iron overload in patients who are receiving long-term blood transfusions in the treatment of chronic anemias.

The FPC had issued a PI in summary proceedings in December 2021; see this Blog here. This finding has been reversed with the present landmark ruling in main proceedings, by an enlarged panel of five judges: The judgment holds that both patents are not infringed, neither literally nor under the DoE. This is what hn ③ is all about.

But let’s take it one step at a time, walking through all headnotes.

Board of appeal decision is proper novuM

Hn ① relates to a procedural matter:

Novartis logo

Let me put this into context first. In litigation at the FPC, parties have two unrestricted opportunities to comment on the merits of the case and, in particular, to introduce new facts into the proceedings. Thereafter, new facts and evidence (nova) can only be introduced under the limited conditions of Art. 229 CPC.

According to the Supreme Court (BGE 146 III 416, r. 4.1; see this Blog here), filing of a new claim set is to be treated in the same way as the submission of other nova. In the underlying case of this Supreme Court judgment, the patentee had partially waived the patent in ex parte proceedings at the Swiss IPI (Art. 24 PatA, comparable to limitation proceedings under Art. 105a EPC at the EPO) only after the closure of the file in co-pending litigation, and was barred from introducing the limited claims into the civil proceedings. The Supreme Court considered the limited claims as an improper novum under Art. 229 para. 1 let. b CPC: Even though the limited claims came into existence only after closure of the file, they could have come into existence earlier, only depending on the decision of the patentee with respect to both timing and extent of the partial waiver.

The present judgment holds that the situation in inter partes opposition / appeal proceedings at the EPO is different. First of all, such proceedings are not initated by the patentee. Second, and even though it is the patentee who formulates the wording of all his auxiliary requests, it is difficult for him to predict which of the auxiliary requests (if any) will survive examination on the merits. Finally, the timing of such inter partes proceedings is determined to large extent by the conduct of the other parties and the procedural calendar of the Board of Appeal. In sum, the FPC holds that the patentee is completely deprived of control over the time and at least partially deprived of control over the version in which the patent is maintained.

Mepha logo

Accordingly, the FPC held that the Board of Appeal decision T 1057/21 to maintain EP 018 in amended form is to be considered as a proper novum under Art. 229 para. 1 let. a CPC.

Let’s have a closer look at T 1057/21. EP 018 was maintained based on auxiliary request #48(!) of 71 auxiliary requests in total. Auxiliary request #48 had been initially filed as auxiliary request #38 in first instance proceedings on 19 February 2021, more than a year before closure of the file in nullity proceedings O2021_004. Patentee re-shuffled his requests later, i.e. the Board of Appeal did not deal with 47 requests before it finally allowed auxiliary request #48. Rather, the patentee made this request his new main request in the oral proceedings before the Boards of Appeal on his own volition, and this new main request was found allowable.

Undoubtedly, it is an uncomfortable situation for patentees to coordinate national nullity proceedings (governed by the strict rules on nova after closure of the file under the CPC) with co-pending central opposition / appeal proceedings at the EPO. Which is why some countries rule out national nullity proceedings while EPO opposition / appeal proceedings are still pending (e.g. Germany; see § 81(2) PatA); not so in Switzerland. The FPC’s approach will surely be appreciated by patentees, if and when it is ultimately approved by the Supreme Court.

burden of proof for technical effects

Hn ② relates to the allocation of the burden of proof for the technical effect(s) of the invention. Now we know:

On a quick read, the hn might suggest that it is up to the patentee to actually prove the technical effect, while the plaintiff in invalidity proceedings can sit back and just grouse that the claimed technical effect has not been proven (yet). But beware: The hn only deals with the allocation of the burden of proof.  It does not say anything about the standard of evidence according to the lex fori. The regular standard of ‘full conviction’ applies in Switzerland, which according to settled case law means that the court should have no serious doubts about the existence of the alleged fact, or that remaining doubts appear to be slight.

Now, what does that mean for assessment of an alleged technical effect of an invention?

The proof does not necessarily have to be provided with empirical data or even clinical studies. If the person skilled in the art can conclude on the basis of scientific or theoretical considerations that the claimed technical effect exists, this is sufficient proof. When assessing the evidence, it may also be taken into account which indications speak against the claimed technical effect. If there are no such indications, the requirements for proving the technical effect are lower.

Tango
it takes two …

Not much of a surprise, isn’t it? In practice, it takes two to tango: The more indications are on the table that the effect is not achieved, the harder it will be for the patentee to actually prove the contrary. Even though the burden of proof lies with the patentee in the first place (which is also the case at the EPO, see e.g. T 1188/00 (hn) and G 2/21 (r. 26)), the whole discussion about whether or not a certain technical effect is achieved is an argumentative arm wrestling.

Numerical ranges / infringemement under the DOE

The judgment holds that both EP 202 and EP 018 as amended in opposition/appeal proceedings are valid (r. 46-67) but not infringed, neither literally nor under the DoE.

The decisive feature in terms of infringement is the amount of deferasirox. The independent claims of both patents require 45-60% by weight of deferasirox. The attacked embodiment contained 64.3% by weight, based on the information given in the product information; r. 68. Even the lowest (disputed) analytical value submitted by the patentee was 60.8%. The judgment holds that even in case of the lower value literal infringement is ruled out. The skilled person will typically consider numerical ranges as more precise and more clear than other verbally described technical claim features (in line with the judgments X ZR 168/00Schneidmesser I (r. II 3 lit. d) and X ZR 135/01Schneidmesser II (r. 4 lit. d) of the German Supreme Court). Accordingly, also values within usual tolerances are not literally covered. The skilled person knows, however, that technical numerical ranges are subject to a measuring error. Therefore, as a rule, the last digit of numerical values corresponds to the measurement accuracy. If no other error limits are indicated (as is the case here), the maximum error for the last digit is estimated from the rounding convention; see e.g. T 175/97 (r. 2.6). The judgment holds that this case law developed for prior art disclosures equally applies to the interpretation of numerical ranges in patent claims. Since the claimed range is given without any decimal digits, the range that is still within the literal scope of protection is ≥44.5% to <60.5% if the normal rounding convention is applied.

Accordingly, and in particular in the absence of any teaching in the patents that would point to a broader interpretation of the claimed range, the judgment holds that neither 64.3% nor 60.8% is literally covered by the claims.

Turning to infringement under the DoE (r. 73 et seqq.), the judgment holds that the variant has the same effect; i.e., Q1 of the three prong test is answered in the affirmative (not much out of the ordinary for generics).

As to the second question re accessibility of the variant (‘Auffindbarkeit’), the judgment holds that there is currently some uncertainty as to whether this should be dealt with in retrospect (i.e., in knowledge of the variant like in the U.K.), or not (like in Germany); r. 77. It is hoped that the Supreme Court clarifies its position when it gets the chance (which is now):

In the case at hand, the FPC held that it does not matter: The same effect of the variant was ‘accessible’ without involvement of an inventive step in any event; r. 78. Thus, Q2 is answered in the affirmative.

Q3 is meant to safeguard the legitimate interest of the public in legal certainty. Here it is (my own understanding / translation):

Does an objective reading of the patent specification lead the skilled person to the conclusion that the patentee has formulated the claim — for whatever reason — so narrowly that he does not claim protection for an embodiment of same effect (Q1) that is ‘accessible’ for the skilled person (Q2)?

The single embodiment of both patents makes use of 56% of deferasirox (para. [0016] von EP 202; para. [0017] von EP 018). In the claims, however, not this specific value is mentioned, but rather a range of 45% to 60%, which is evidently placed around this specific value. Under these circumstances, the judgment holds (r. 80) that the skilled person assumes that the patentee has specifically and bindingly placed the claimed range around the specific value, i.e. the patentee has deliberately decided to specify the invention to be within this range and thus to dispense with the scope of protection lying outside this range. Under the circumstance, the judgment holds that there is no room to extend this range under the DoE by taking into account generally accepted tolerances going beyond it, because the skilled person may in good faith assume that the patentee has already taken into account possible tolerances when defining the range. In passing, the judgment notes that the range is not symmetrically placed around the specific value; it rather is narrower above than below the specific value. This even underlines the importance of the upper limit. Thus, infringement under the DoE was denied.

This outcome is essentially in line with the U.K. judgment [2022] EWHC 2847 (Pat), para. 232 (emphasis added):

I have found that in the present case the inventive concept of the Patents would be seen by the skilled team as requiring strict compliance with the 45-60% range of deferasirox. On that basis a variant will not be substantially the same as the inventive concept unless it strictly complies. I need not investigate with precision what compliance means numerically as matter of general principle. In my judgment Teva DFX does not comply.

In the Swiss case, Novartis appealed to the Supreme Court. The appeal is still pending at the time of writing.

In accordance with the outcome in main proceedings, the PI was lifted with separate order.

UPDATE 28 Sept. 2023: Novartis’ appeal has been dismissed by the Supreme Court; the decision is final.

✍ MW

BIBLIOGRAPHY

Case Nos. O2021_004, O2021_005 | Judgment of 20 April 2023 | ‘Deferasirox II’

Mepha Pharma AG
./.
Novartis AG

Panel of Judges:

    • Dr. Mark SCHWEIZER
    • Dr. Tobias BREMI
    • Dr. Rudolf RENTSCH
    • Prisca von BALLLMOOS
    • Marco ZARDI

Judge-rapporteur:

    • Dr. Tobias BREMI

Court Clerk:

    • Sven BUCHER

Representative(s) of Plaintiff:

Representative(s) of Defendant:

JUDGMENT
of 20 April 2023 in main proceedings

ORDER
of 12 May 2023 rectifying / supplementing the judgment

JUDGMENT
of 15 December 2021 in summary proceedings

PATENTS IN SUIT

EP 2 964 202 B1:

See EPO Register and Swissreg for bibliographic details.

EP 2 964 202 B2:

See EPO Register and Swissreg for further information.

EP 3 124 018 B1:

See EPO Register and Swissreg for further information.

EP 3 124 018 B2:

Not yet available. Currently being dealt with by the OD again, after T 1057/21,  to compile the set of documents for maintenance of the patent in amended form, based on the following claims:

See EPO Register and Swissreg for further information.

EPO BOARD OF APPEAL DECISIONS

T 0526/21 re EP 202

T 1057/21 re EP 018

BE ON THE KNOW

Enter your name and email address below to get notified of new posts by email.

Equivalency in case of numerical ranges

Reading time: 9 minutes

Case No. S2021_005:
Order of 19 August 2021 | ‘Deferasirox-Mepha’
Decision of 15 December 2021 | ‘Deferasirox-Mepha’

DISCLOSURE NOTICE: Hepp Wenger Ryffel is involved in this matter on behalf of Defendant.
deferasirox

This case is about deferasirox, an iron chelating agent. It is mainly used to reduce iron overload in patients who are receiving long-term blood transfusions in the treatment of chronic anemias.

Novartis held a compound patent that already expired in June 2017; EP 914 118 B1 (EPO Register and Swissreg). A subsequent SPC expired in May 2021; C00914118/01.

Exjade

Initially, Novartis had sold deferasirox as dispersible tablets under the trademark Exjade®. However, the dispersible tablets have been taken off the market in October 2019. They were replaced by Novartis’ Jadenu®, i.e. swallowable tablets of deferasirox.

Mepha’s Deferasirox-Mepha had been approved by Swissmedic on 24 August 2020, and Mepha lodged nullity proceedings against EP 2 964 202 B1 (EPO Register and Swissreg) and EP 3 124 018 B1 (EPO Register and Swissreg) on 19 April 2021 (case no. O2021_004). The patents in suit are currently also challenged in opposition/appeal proceedings at the EPO; both patents were maintained in limited form in first instance, with their respective main claims as follows:

In response to the nullity attack, Novartis counter-claimed for infringement on 29 June 2021 (case no. O2021_005).

Novartis logo

Shortly thereafter, Novartis requested a preliminary injunction against Deferasirox-Mepha on 11 August 2021, based on EP 202 and EP 018. A request for ex parte preliminary injunctive relief was dismissed with order of 19 August 2021, for lack of particular urgency.

The summary judgment after hearing the defendant holds that both patents do not suffer from unallowable amendments, cf. ¶¶25-27. Novelty had not been challenged, and inventive step over WO 2004/035026 A1, WO 2007/045445 A1, WO 2009/067557 A1 and WO 2010/035282 A1 is acknowledged in the summary judgment, cf. ¶¶29 et seqq..

Mepha logo

With respect to infringement, it is important to recall that both patents require deferasirox to be present in an amount «from 45% to 60% by weight»; the allegedly infringing embodiment contained 64.3% by weight of deferasirox.

Novartis essentially argued that numerical values are not sacrosanct but need to be construed like any other feature, and that in particular in view of typical tolerances a value of 64.5% by weight would still be literally covered by the range of from 45% to 60% by weight. The FPC did not buy into that, in no uncertain terms:

Now, moving on to the assesement of infringement under the DoE, here is the current Swiss three-step questionnaire (as a reminder), i.e. assessment of i) same effect(s); ii) accessibility; iii) equality / equal value:

Q1: Same effect?

The judgment first elaborates on two essential issues in the assessment of Q1, i.e. the effect(s) of what; and just one, some or all effect(s)?

As to the first issue, the Supreme Court has caused some confusion amongst practitioners with a remark in BGE 143 III 666Pemetrexed that, at least at the face of it, might suggest that the effect of the whole claimed subject-matter is decisive, not only the effect of the replaced feature (r. 5.3.3, emphasis added):

Vielmehr muss die abgewandelte Ausführungsform alle diejenigen Wirkungen erzielen, die nach dem Verständnis des Fachmanns mit den einzelnen technischen Merkmalen des Patentanspruchs für sich und in ihrem Zusammenwirken erzielt werden sollen.

However, the Supreme Court has also explicity referred to the case-law of the German Supreme Court in the Pemetrexed case, and the German case-law undoubtedly focusses on the effect(s) of the replaced feature only. In sum, the FPC holds that it cannot see a clear deviation of the Supreme Court from the generally applied focus on the effect(s) of the replaced feature. (I hope the Supreme Court will clarify its position at the next possible occasion.)

On the second issue, i.e. which / how many effects of the replaced feature one has to look at, the judgment holds that only the desired effects of the replaced features matter (referred to as ‘erfindungsgewollte’ or ‘patentgewollte’ effects) — but that all these desired effects have to be considered (see emphasis in the above citation).

The judgment holds hat the same effect in terms of bioequivalency and swallowability suffices in the case at hand. The mere fact that plaintiff failed to establish the further desired effects of the replaced feature which are outlined in the patents in suit is held to be irrelevant because defendant could not prove that the challenged embodiments do not achieve the corresponding effects either.

Accordingly, Q1 is answered in the affirmative (same effect(s) given).

Q2: Accessibility?

Note that Q2 asks for ‘accessiblity’ of the same effect in retrospect, i.e. when the skilled person is confronted with both the claimed feature and the variant with the replacing feature: In knowledge of both, is the same effect ‘accessible’ for the skilled person?

This approach is in line with e.g. the UK Supreme Court in [2017] UKSC 48 (cf. ¶62), but markedly different to the approach of the German Supreme Court which excludes non-obvious variants from infringement under the DoE, i.e. variants that can only be found with inventive activity.

The FPC notes that, again, the Supreme Court judgment Pemetrexed (BGE 143 III 666) is not fully clear.  While the earlier Supreme Court judgment BGE 142 III 772 — Urinalventil explicitly held that an inventive variant (where exactly the replacing feature is inventive over the replaced one) can constitute an infringement under the DoE (cf. r. 6.4), the exact opposite is held in the Pemetrexed case (cf. r. 5.4.1):

The skilled person must, on the basis of his general knowledge, be stimulated by the patented invention to make the variation; if the variation is itself based on an inventive step, accessibility is ruled out.

Notably however, the Supreme Court in Pemetrexed did not refer to the Urinalventil case anymore, but rather to a pretty old case that still applied a ‘gist of the invention’ approach — which is rightly considered overcome today. (Again, I do hope that the Supreme Court will clarify its position at the next possible occasion.)

Finally, the FPC holds that the current wording of Q2 may be criticized with good cause (sic!), but that it surely is not totally off-beat since it is also used by the UK Supreme Court. The FPC did not consider it appropriate to deviate from the current wording of Q2 in summary proceedings; this shall be subject to main proceedings at the FPC — and ultimately the Supreme Court (again).

In application of the current (retrospective) wording of Q2 (cf. p. 52, last para.), the judgment holds that Q2 is to be answered in the affirmative (accessibility given). However, I’m afraid the way the conclusion (and the analysis two paragraphs earlier) is phrased might well add to the confusion:

Q2 as it stands is not about accessibility of the replaced feature per se, but rather about accessibility of the same effect of the replaced feature. (‘Accessibility’ (Auffindbarkeit) is a very misleading term in the context of Q2 anyway, in my perception; ‘recognisability’ (Erkennbarkeit) would make more sense to me.)

Q3: Equal value?

The judgment holds that tolerances beyond a claimed range which are generally accepted in a certain field are covered by the claimed range, under the DoE:

In view of the evidence on file (a meanwhile replaced EMEA ‘Note for Guidance on Manufacture of the finished dosage form‘ (1995) and European Pharmacopoiea 10.0, 2.9.5), the judgment holds that an amount of 64.4% by weight of the active ingredient is well within the generally accepted tolerance (66% by weight); cf. ¶¶59-61. Q3 was thus answered in the negative.

Accordingly, the summary judgment holds that the patents in suit are infringed under the DoE. Interim injunctive relief was thus granted, but not to the extent of storage (‘Lagern’) and possession (‘Besitzen’); cf. p 60 (don’t be misled by the wrong marginal numbers). Further, defendant was ordered to recall the product from commercial customers within three days of receipt of the judgment; 24 hours as requested were considered too short.

As mentioned above, main proceedings are already pending.

Reported by Martin WILMING

BIBLIOGRAPHY

Case No. S2021_005:
Order of 19 August 2021 | ‘Deferasirox-Mepha’
Decision of 15 December 2021 | ‘Deferasirox-Mepha’

Novartis AG
./.
Mepha Pharma AG

Panel of Judges:

    • Dr. Mark SCHWEIZER
    • Dr. Tobias BREMI
    • Marco ZARDI

Judge-rapporteur:

    • Dr. Tobias BREMI

Court Clerk:

    • Susanne ANDERHALDEN

Representative(s) of Plaintiff:

Representative(s) of Defendant:

    • Andrea MONDINI (TIMES)
    • Dr. Andreas WELCH (Hepp Wenger Ryffel), assisting in patent matters

ORDER IN FULL

JUDGMENT IN FULL

PATENTS IN SUIT

BE ON THE KNOW

Enter your name and email address below to get notified of new posts by email.