Non-reproducible prior art post-G 1/23: Novelty versus inventive step (T 1044/23)

The Board of Appeal decision in T 1044/23 provides us with a first look at how the principles of G 1/23 will be applied to black-box prior art. The decision confirms that whilst novelty attacks based on prior use and sale of a non-reproducible product are possible, the path to an inventive step attack based on such disclosures may still be challenging in the face of the practical realities of well-protected trade secrets. G 1/23 and its interpretation by the Boards of Appeal highlights the potential important interplay between patents and trade secrets in an effective overarching IP strategy. 

Legal background: The shift from G 1/92 to G 1/23 on enablement

The recent decision of the Enlarged Board of Appeal (EBA) in G 1/23 revisited a question previously considered in G 1/92, relating to what constitutes prior art with respect to a novelty and inventive step assessment. According to Article 54(2) EPC:

"The state of the art shall be held to comprise everything made available to the public by means of a written description, by oral description, by use, or in any other way, before the date of filing of the European patent application." 

Black-box prior art

Whilst Article 54(2) EPC therefore broadly captures all public disclosures as being prior art, the EBA in G 1/92 introduced a critical enablement requirement regarding "availability". Specifically, in G 1/92, the EBA found that a product on the market was not legally "available" under Article 54(2) EPC if the skilled person could not reproduce it without undue burden. In other words, the fact that a product cannot be reproduced meant that the product itself could not be considered prior art. 

The peculiar legal fiction established by G 1/92, whereby a product could be purchased or otherwise made available to the public, yet legally ignored in an assessment of patentability, required certain mental gymnastics to understand. G 1/23 effectively did away with this odd position. In G 1/23, the EBA found that a commercial product cannot be excluded from the state of the art solely due to non-reproducibility (IPKat). As confirmed in the headnote to G 1/23:

"A product put on the market before the date of filing of a European patent application cannot be excluded from the state of the art within the meaning of Article 54(2) EPC for the sole reason that its composition or internal structure could not be analysed and reproduced by the skilled person before that date."

Case summary of T 1044/23: Polyethylene composition and prior use

We now have our first Board of Appeal to apply the order of G 1/23. The patent in question in T 1044/23 (EP 3161066 B1) related to a polyethylene composition suitable for injection moulding. The main issue on appeal was whether the prior sale of certain commercial resins, specifically products sold under the names DMDA-8904 NT 7 and DMDA-8907 NT 7, anticipated or rendered obvious the claimed invention. 

The invention aimed to solve the problem of providing a polyethylene composition with a specific balance of properties, including density, melt flow indices, and a specific polydispersity index known as the "ER value". At the first instance, the Opposition Division, citing G 1/92, rejected the opposition, reasoning that the commercial resins DMDA-8904 and DMDA-8907 were not prior art under Article 54(2) EPC because their manufacture was not enabled (OD decision, para. 19.5). The OD thus took the pre-G 1/23 approach to prior use, namely if the skilled person could not reproduce the product without undue burden, it was not "available" to the public (IPKat). 

On appeal, the Opponent argued that, following G 1/23, the reproducibility requirement was no longer a valid criterion to exclude a commercial product from the state of the art. The Opponent submitted evidence, including invoices and certificates of analysis, to prove that the resins had been sold and delivered before the priority date. The key question for the Board of Appeal was thus whether these "black box" resins, whose precise manufacturing process (catalyst and conditions) was seemingly a trade secret, could be cited against the patent for novelty and inventive step following G 1/23.

Novelty assessment: Applying G 1/23 to non-reproducible products

Citing the headnote of G 1/23, the Board of Appeal confirmed that a product put on the market cannot now be excluded from the state of the art solely because its composition or internal structure cannot be analysed and reproduced. The Board of Appeal concluded that the reproducibility requirement is thus no longer a valid condition to assess whether a product is state of the art within the meaning of Article 54(2) EPC (r. 2.4.3). 

Having established and confirmed the new legal test for prior use under G 1/23, the Board of Appeal turned to the facts of the case. The Board of Appeal was convinced that the documents submitted by the opponent proved the public availability of the resins (r. 2.5.6). Furthermore, for the Board of Appeal, the evidence showed that the composition of these commercial resins fell squarely within the scope of claim 1 as granted (r. 2.6.1). The Board of Appeal thus concluded that the subject-matter of granted claim 1 lacked novelty with respect to these resins (r. 2.7).

Non-reproducible products in the problem-solution approach

The case in T 1044/23 then moved to an auxiliary request, in which the Patentee had amended the density range to exclude the specific density of the prior art DMDA resins. This was considered to restore novelty but was objected to for lack of inventive step (Article 56 EPC).

The Patentee argued that because the methods to prepare the DMDA resins were not publicly available and could not be reproduced, the prior art resins were not a suitable starting point for assessing inventive step. The Patentee essentially argued that one cannot rationally start from a product if one does not know how it was created.

However, the Board of Appeal was not convinced by this argument. Referring again to G 1/23, the Board of Appeal noted that the EBA had explicitly stated that "depending on the circumstances, also a non-reproducible product may be considered to represent the closest prior art"  (r. 5.1.6). The Board of Appeal concluded therefore, that even if the skilled person would not know how to reproduce these resins, this was not a sufficient reason to disregard them as the closest prior art. The Board of Appeal reasoned that because the resins were commercially available, a skilled person could use them as such or attempt to modify them. The Board of Appeal drew a parallel to the Coca-Cola example in G 1/23, noting that the inability to reproduce the original recipe does not stop a skilled person from trying to improve the product. 

Critically, the Board of Appeal distinguished a non-reproducible commercial product from a "defective or speculative disclosure," noting that the former is a "physical, tangible product" (r. 5.1.9). This is an important distinction, which means that following G 1/23 it is still possible to dismiss non-enabling speculation in the prior art, which is particularly important in the pharma industry where hypothesizing is often easy, but execution is hard. 

The Board of Appeal thus concluded that the commercial resins could be selected as the closest prior art. As in the case underlying G 1/23, the decision in T 1044/23 therefore confirms that the "starting point" in the problem-solution approach can be a product you can buy but cannot make.

The surviving value of trade secrets

Selecting a starting point is just the first step in the problem-solution approach. There remains the question of whether the invention is obvious in view of the closest prior art. The problem to be solved in this case was formulated by the Board of Appeal as the provision of an alternative polyethylene composition. The question was whether it would have been obvious for the skilled person to modify the DMDA resins to arrive at the claimed density range whilst maintaining the other parametric requirements (MIE, MIF, and ER values).

The Opponent argued that it was common general knowledge to adjust density by varying comonomer content and that the general process type (UNIPOL gas-phase) was known. They argued that the skilled person could reproduce the resins or obtain similar ones and then adjust the density.

The Board of Appeal, on the other hand, saw a distinction between general process knowledge and the specific know-how required to replicate a complex commercial polymer. The Board of Appeal concurred with the Patentee that the exact processes were likely confidential know-how: "it has not been shown that, by following these indications and with only routine experimentation, the skilled person would have been able to reproduce the DMDA resins or obtain materials having comparable properties" (r. 5.2.3). 

The Board of Appeal thus concluded that whether the skilled person could achieve the claimed composition without undue burden "remains a matter of conjecture". Consequently, the invention defined in the auxiliary request was not shown to be obvious. The appeal was therefore dismissed (in part regarding the auxiliary request) and the case remitted with the order to maintain the patent on the basis of the auxiliary request.

Final thoughts

The decision in T 1044/23 illustrates some of the practical boundaries of the EBA decision in G 1/23 with respect to novelty and inventive step. For a novelty assessment, if the product exists and matches the claims, novelty is destroyed with respect to the product. It does not matter that the product cannot be reproduced by a skilled person. 

For inventive step, G 1/23 allows the use of non-reproducible prior art as the starting point for inventive step. However, trade secrets may continue to provide considerable protection if they prevent the skilled person from working out how to arrive at the claimed invention. As we saw in the referring case to G 1/23, the important question is whether the skilled person can arrive at the invention as claimed and not the particular cited prior art product. If an opposing party cannot explain, in view of the protected trade secrets, how the skilled person would modify a specific, non-reproducible product to reach the invention, an inventive step attack may fail for lack of a reasonable expectation of success or enabling disclosure for the modification. T 1044/23 confirms that whilst G 1/23 removed the "legal fiction" that non-reproducible products don't exist, it did not remove the practical reality that you cannot modify what you do not understand. To this Kat, G 1/23 is one of the EBA's clearest and sensible decisions in recent years. 

All of this highlights the importance of an effective and thoughtful trade secret policy. In fields in which products must necessarily be placed early on the market at an early stage, the question of what processes and information would enable a skilled person to arrive at new commercially important developments, and whether these may be protected by trade secrets, may be a highly valuable strand in an effective IP strategy. 

Further reading

• New EBA referral: When is prior use of a product excluded from the prior art for lack of enablement? (Jan 2023)
• Non-reproducible commercial products are prior art (G 1/23) (Nov 2025)
• [Guest Post] Has G1/23 harmoniously recast the European law of enablement? (Nov 2025)
• epi podcast: G 1/23: What do we know?