Clinical trial protocols as prior art at the EPO

In a claim directed to a product for use in a medical treatment, the therapeutic effect is considered a functional, technical feature of the claim.  For sufficiency of disclosure, the application as filed must disclose the suitability of the product for that claimed medical use, unless this is already known to the skilled person at the relevant date (Case Law, II.C.7.2.2.).

The suitability of the product for a further medical use is often established by the results of clinical trials.  However, in the clinical trials process, a trial protocol is typically published in advance of the trial being conducted.  That means the product and the therapeutic use are disclosed, but without any results.

Accordingly, since the application is only filed after the results of the trial are obtained, the protocol itself is prior art.  The protocol is generally not novelty-destroying, because the therapeutic use has not been shown (the prior art is not enabled in the same way a patent application would be insufficient without any evidence of the therapeutic use).  However, inventive step remains a key issue.  This scenario has led to a number of interesting cases at the EPO, culminating in a recent case relating to cabazitaxel (T0136/24), which has been the subject of much discussion.

The early development of the case law

T2506/12-3.3.07 (pegylated liposomal doxorubicin and ecteinascidin)

The board confirmed that the disclosure of an ongoing clinical trial did not constitute a disclosure of the therapeutic effect.  Nevertheless, it was emphasised that the success of the trial was obvious, because clinical trials involving human subjects are not undertaken on a “try-and-see” basis, but rather on the basis of favourable existing scientific data, driven by ethical and economic considerations (T2506/12, r. 3.10).  The board further stressed that the inherent uncertainty of clinical trials does not render their outcome inventive (T2506/12, r. 3.12.2).

T0239/16-3.3.01 (zoledronic acid)

This reasoning was followed in T0239/16, where a clinical trial protocol was accepted as the closest prior art for the claimed dosage regimen of zoledronic acid.  Given that the relevant class of active agents was already known to be effective for the claimed indication, the board held that authorisation of the clinical trial itself generally implied a reasonable expectation of success.  This conclusion was based on the understanding that clinical trials are grounded in preclinical data and that their approval reflects the ethical and economic considerations that a benefit will arise with “reasonable certainty” (T0239/16, r. 6.5 and 6.6).

T1123/16-3.3.04 (eosinophilic bronchitis)

The board maintained the presumption of a reasonable expectation of success, absent an expectation of failure (T1123/16, r. 11).

T0096/20-3.3.04 (anti-C5 antibody)

T0096/20 stressed the prevailing position and held that the announcement of a detailed safety and efficacy clinical trial protocol for a specific therapeutic and disease provided the skilled person with a reasonable expectation of success, unless there was evidence to the contrary in the prior art (T0096/20, r. 8 and 9).

This decision represented the “high point” of assumed obviousness.

T1806/18-3.3.01 (nilotinib in apple sauce)

The fact pattern in this case was different in that the prior art was a paediatric investigation plan, rather than a protocol for a clinical trial.  The board held that there was no evidence that a PIP was based on the same rigorous scientific assessment as a clinical trial protocol.  When starting from the PIP, they stated that the announcement of a clinical study “does not automatically mean that its outcome was predictable and that a reasonable expectation of success had to be acknowledged. Whether this is indeed so, depends on the facts and circumstances of each case” (T1806/18, r. 7.21).  In this case, the unpredictable food effect on the bioavailability of the drug in question, together with safety concerns, meant that a successful outcome could not be assumed simply because a PIP had been approved.

T3165/19-3.3.08 (PCSK9 antibody)

The issue here was that additional prior art showed that “for this patient group [inadequate control of other atherogenic lipoproteins despite treatment with a maximum tolerated daily dose statin therapy]” there “was no clear link between the [pharmacological effect – reduced LDL-C levels] and [the claimed indication – reduced cardiovascular risk]” (T3165/19, r. 17).  Based on this additional information, the board held that “the skilled person could not have had more than a hope that the treatment assessed in the phase 3 clinical trial disclosed in document D4 would succeed in achieving its goal of reducing cardiovascular risk in the treated patients” (T3165/19, r. 23).  Thus, this case followed the established approach, but like T1806/18, there were additional factors pointing away from an expectation of success.

Some recent cases from board 3.3.07

T1437/21-3.3.07 (empagliflozin)

By the time of this decision, the composition of board 3.3.07 had completely changed from the board that decided T2506/12.

T1437/21 related to the second medical use of a known drug, empagliflozin, for treating diabetes in patients with “moderate renal impairment” – a subgroup associated with therapeutic uncertainty because the active, empagliflozin, acts via the kidneys.

The prior art was the patentee’s own press release reporting successful phase III data.  The report said that patients were included with mild, moderate and severe renal impairment.  But positive results were stated only for all patients, and not individualised for each patient group.  The board held that neither the existence of a phase III trial nor positive topline results created a reasonable expectation of success in the moderate subgroup.  And that, on the facts of the case, there was no reasonable expectation of success in patients with moderate renal impairment.

T1941/21-3.3.07 (tauroursodeoxycholic acid)

The prior art showed that most drugs which had been tested in clinical trials for the same medical indication had failed, including compounds with the same pharmacological properties as the claimed active.  Based on this, the board held that the skilled person would have at best “a hope of success…but more likely an expectation of failure” (T1941/21, r. 1.5.1), despite the announcement of the claimed therapeutic in a phase II clinical study.

The approach of board 3.3.04 as set out in T0136/24-3.3.04 (cabazitaxel)

The composition of board 3.3.04 is now different to the board that decided T0096/20.  Although interestingly, the chair of board 3.3.04 was the rapporteur in T0239/16.

The opposed patent claimed cabazitaxel co-administered with prednisone for treating metastatic castration-resistant prostate cancer after prior docetaxel therapy.  The key prior art was a phase III clinical trial protocol that compared cabazitaxel and prednisone to mitoxantrone and prednisone in the claimed setting.

The opponents relied on earlier case law (notably T2506/12, T0239/16, T1123/16 and T0096/20) to argue that “the announcement of a clinical study established a legal presumption of reasonable expectation of success, with the consequence that a patent proprietor had the burden of rebutting such a legal presumption by showing that the skilled person would have been dissuaded by the prior art from following the study protocol with a reasonable expectation of success and so achieve its inevitable outcome” (T0136/24, r. 7.11).

The board dismissed this legal presumption approach, highlighting the case law on balancing positive and negative pointers (T0136/24, r. 7.14 and 7.14.1).  On that basis, the board held that “the question of whether there was a reasonable expectation of success must be answered on the basis of the specific circumstances of the case”.  This required an evaluation of all the facts available at the relevant date of the contested patent, including inter alia the nature of the active agent under investigation, and how much was known about the clinical efficacy, safety and potential toxicities of the active agent(s) in the pertinent therapeutic indication (T0136/24, r. 7.14.4).

The board also stated that the question of obviousness must be answered by considering whether there was a reasonable expectation that the trial’s primary endpoint (improving overall survival) would be met, even though the claim was not limited to treatment assessed by that parameter (T0136/24, r. 7.13).

The board took the view that the position they had taken was consistent with earlier decisions of the boards, which generally required a balancing of the expectation offered by the protocol itself, and the positive and negative pointers in the prior art.  However, the board explicitly stated that they did not follow the more categorical finding in T0096/20 that there had to be an expectation of success absent factors pointing away (T0136/24, r. 7.14.10).

Applied to cabazitaxel, the fact-specific analysis found limited clinical data in the claimed setting and a lack of preclinical support.  On this basis, the board found that the skilled person had at best a hope that cabazitaxel would improve overall survival in the claimed patient group, not a reasonable expectation of success.

Interestingly, the local division of the UPC in Munich found the corresponding patent to lack an inventive step.  The court explicitly disagreed with the reasoning in T0136/24, insofar as the skilled person would have to reasonably expect that the primary endpoint of the clinical trial (overall survival) would be met (UPC decision, page 33, first sentence; c.f. T0136/24, r. 7.13): “However, it is important to note that the question of reasonable expectation of success of the approach disclosed in the [trial] documents, in terms of assessing an inventive step, is not the same as the question of whether the [trial] will meet its primary endpoint.”

The UPC also disagreed on the facts, finding that the balance of positive and negative pointers suggested a reasonable expectation of success.

Final thoughts

Changes in emphasis have been subtle over the years, but the present situation from the two chemical boards who typically handle such cases, board 3.3.04 and 3.3.07, seems fairly well aligned.  They are taking an approach which reduces the weight placed on the trial protocol itself, and leaves the parties to fight the battle on the balance between positive and negative pointers, i.e. building a case on whether or not there was a reasonable expectation of a positive outcome for the trial based on the available evidence, with minimal reference to the protocol itself.

About the author

Dr Richard Gillard specialises in opposition/appeal work at the EPO and was involved in T0239/16, T1806/18, T1437/21 and T0136/24.

For more detailed advice in relation to any of the issues discussed above, or for advice relating to other matters regarding European practice, please do not hesitate to get in contact with your usual E+F advisor or email us at elkfife@elkfife.com.