The 1973 European Patent Convention (EPC) harmonised patent law throughout the contracting states and was a very significant achievement. The contracting states include all the member states of the EU and some other European countries.
When it was signed, it was hoped that the EPC would be interpreted uniformly, both by the European Patent Office (EPO), its central registration body and by the national courts of the contracting states, each of which has sole jurisdiction over infringement and is entitled to consider the validity of their resulting national patents.
Significant differences remain in the procedural laws, both between the national courts of the different contracting states, and between those national courts and the EPO. These procedural differences, particularly regarding the treatment of evidence and witnesses, can lead to the EPO and national courts coming to different conclusions regarding the patentability and validity of effectively the same European patent.
This has led to some outspoken commentary in two recent judgments of the English Court of Appeal. These cases, each concerning high-value pharmaceutical patents, highlight the rocky road ahead, both for a unified pan-European patent and for a single patent court competent to determine questions of patent enforcement and validity across all member states of the EU. This article explores some of the issues and the relevant UK cases, against the backdrop of the patenting priorities of Europe’s pharmaceutical sector.
UK approach to substantive patent law
UK courts have long sought to align principles of UK patent law with EPO jurisprudence. In 1996 the House of Lords held in Merrell Dow Pharmaceuticals Inc v HN Norton & Co Ltd  that in construing the UK Patents Act 1977, UK courts must have regard to decisions of the Technical Boards of Appeal (TBAs) of the EPO on the construction of the EPC since, although not strictly binding, they are of ‘great persuasive authority’. The House of Lords added that it would be ‘highly undesirable’ for the provisions of the EPC to be construed differently by the EPO and the national courts of a contracting state.
Recent case law: Actavis UK Ltd v Merck & Co Inc 
Subsequent UK case law has strongly supported the Merrell Dow approach. Actavis v Merck is a recent example. Here, the Court of Appeal held that it could (and normally would) overrule its own earlier decisions to follow the settled practice of the EPO. The effect was to create a new exception to the longstanding rule in Young v Bristol Aeroplane  that the Court of Appeal is bound by its own previous decisions.
Actavis v Merck questioned whether a patent could be granted for the use of a known pharmaceutical for the preparation of a medicament for a new dosage regime. All patentable inventions must be new, but no patent can be granted for a method of medical treatment. The Court of Appeal held that a product for a new, non-obvious, dosage regime could be patentable, following the settled view of the EPO in Esai, even though this seemed to be contrary to their own earlier decision in Bristol-Myers Squibb (BMS) Company v Baker Norton Pharmaceuticals Inc & anor .
In BMS the Court of Appeal concluded that the patent was invalid as it contained nothing new (and consisted of a method of treatment). It also appeared to have concluded that to be valid, the patent had to claim the use of a known compound to treat a disorder that was ‘distinct and different’ from the previously known use. Following a detailed examination of this point, the Court of Appeal in Actavis v Merck concluded that there was no clear ratio to this effect in BMS. This paved the way for the Court of Appeal to follow the EPO case law.
The Court of Appeal pointed to the European patent system and the importance of EPO decisions in further concluding that, in any event, it was entitled to follow the EPO because:
‘If the UK is out of line, it will either be going too far or not far enough; either recognising as valid a patent which the EPO has held should not as a matter of law be granted… or holding invalid patents which the EPO considers are valid. Either situation would be bad for the European market.’
Of course, there will be a handful of cases where it is difficult to determine whether the EPO has a taken a ‘settled’ view on a particular legal issue, especially as it has no binding rules of precedent. (The patentability of business methods and computer programs immediately spring to mind as areas where there is continuing legal uncertainty at the EPO.) In discussing this problem, the Court of Appeal (adopting the language of Merck’s counsel), likened the EPO to a commodore steering the 34 EPC member ships:
‘In the unlikely event that we are convinced that the commodore is steering the convoy towards the rocks we can steer our ship away. Technically we are not in the same position as we are in the case of decisions of the European Court of Justice… And of course if there is no clear message from the commodore or he gives mixed messages we must decide our own course anyway.’
Thus, where there is any doubt as to the EPO’s position, there is a clear risk of conflicting decisions on questions of law. For the most part though, the English courts will follow the EPO.
two UK judgments
Actavis UK Ltd v Novartis AG: procedural differences
A patent is invalid if it does not involve an inventive step. A step will not be inventive if it is obvious. The EPO has a settled ‘problem, solution’ approach (PSA) to questions of obviousness. What was the problem and how was it solved? Was that solution obvious? In Actavis v Novartis the Court of Appeal contrasted the UK’s alternative ‘structured’ approach to the same issue.
The two approaches are summarised in the adjacent table. Significantly, Jacob LJ indicated that the differences between the two methods are only procedural. The legal test for obviousness is set out in article 56 of the EPC and ‘any paraphrase or other test is only an aid to answering the statutory question’.
The four-step UK approach is set out in Pozzoli SPA v BDMO SA & ors . According to the Court of Appeal, the first three steps merely orient the tribunal properly. Step 4 is the key step and poses the fundamental statutory question.
In Actavis v Novartis Jacob LJ took the opportunity to comment on the PSA. Step one was a useful way of containing the very large number of pieces of prior art relied on by opponents before the EPO, since EPO procedure has nothing to prevent them relying on as many as they wish. This is in stark contrast to the UK, where practitioners confine themselves to the best pieces of prior art, as they know that any profligate citation of prior art may be met with an adverse costs order for wasteful conduct. The main difficulty with step one of the PSA is that it can swiftly develop into a satellite dispute as to which is the closest prior art.
The next difficulty arises from the ‘objective technical problem’, ie the problem posed by the nearest prior art, which the patent purportedly seeks to solve. In practice, as the EPO recognises, the inventor may start, not from the closest prior art but from some larger or different problem. Thus the EPO sometimes has to ‘reformulate’ the problem that the patentee thought they were tackling. Jacob LJ thought that this ‘reformulation’ was perhaps the weakest element of the PSA and that the UK approach was preferable, as it left the question of obviousness wide open. There is no attempt to force it into a sometimes artificial problem/solution mould.
Jacob LJ added that the PSA is not always suitable for cases where the invention tackles a known problem that has been ‘put up with’ for years. Nor does it assist with cases involving purely arbitrary limitations that, though not obvious, are also clearly not inventive, such as specifying the colour of a machine or the diameter of a dinner plate.
The Court of Appeal suggested that the open-ended question posed by Pozzoli step 4 avoids such problems and allows for a more flexible, multifactorial assessment of any given case, such as:
- the motive to find a solution to a problem addressed in the patent;
- the number and extent of possible avenues of research;
- the effort involved in pursuing them; and
- the expectation of success.
In pointing out the limitations of the PSA, Jacob LJ implied that Pozzoli provides a sharper instrument for dissecting questions of obviousness. He was, however, quick to add that the PSA works well when there is no need to reformulate the problem and it acts as useful tool for an examining office that needs a common structured approach.
On the facts of Actavis v Novartis, Jacob LJ concluded that both the UK and the EPO approaches led to the same outcome. In some cases, however, procedural differences between these two tribunals can lead to diametrically opposing results.
Eli Lilly & Company v Human Genome Sciences Inc (HGS) 
In HGS the Court of Appeal reached a different conclusion from a TBA on the patentability by HGS of a nucleotide and amino acid sequence of a novel member of the anti-tumour ligand family. This was notwithstanding that the two tribunals co-operated during the proceedings to minimise the risk of unnecessary duplicate proceedings.
Jacob LJ (in another leading judgment) attributed the contradictory outcomes to procedural differences between UK courts and the EPO. While acknowledging that the UK courts will follow any principle of law clearly laid down by the EPO, Jacob LJ did not see why the UK court’s intense fact-finding and assessment process should give way to the findings of fact by the EPO. In his view, national courts are entitled to differ from the EPO in their findings of fact, particularly where cases involve a question of degree over which judges can legitimately differ. The same principle has been upheld several times by the House of Lords in earlier cases.1
HGS is interesting because of Jacob LJ’s analysis of the key procedural differences between the EPO and UK courts. UK trials involve intensive investigation and testing of evidence. Each party must marshal all its evidence and arguments for the trial at first instance, and must disclose all internal documents relevant to the issues in the case.
Both factual and expert evidence are rigorously tested by cross-examination. In the words of Jacob LJ, expert evidence will be ‘probed without remorse’ before a judge who will have already developed a good understanding of the technical subject matter. Many English trial judges will have a science degree and many years’ experience of practising patent law.
The combined effect is that parties must make their best efforts for the first instance trial and put their cards on the table (which in itself causes cases to settle). In addition, the UK Court of Appeal gives considerable deference to the findings of fact at first instance. There are no new witnesses, expert opinions or other new evidential matters on appeal, save in very exceptional circumstances. In short, first instance proceedings in England are ‘for real’:
‘Not only must a party put its cards on the table prior to the trial, it must find and play all its best cards at the trial.’
EPO opposition proceedings are quite different. As they are part of the grant process, it is possible to have on the panel the examiner who allowed the grant of the patent that is under challenge. The court also noted that the EPO has much less room for the testing of evidence in comparison to UK procedure:
‘It cannot, need not and does not investigate matters affecting validity as profoundly as a national court can.’
In practice there is no cross-examination or compulsory disclosure. There is, however, far greater scope for the admission of new evidence on appeal. (In HGS the TBA admitted a further 700 pages of evidence from the patentee just three weeks before the hearing.) Jacob LJ thought it inevitable in these circumstances that the benefit of the doubt would be accorded to the patentee.
Jacob LJ considered it to be well-established that national courts can differ from the TBAs on findings of fact and questions of degree. The two fora would ‘inevitably’ differ in their conclusions from time to time. In HGS, therefore, the Court of Appeal did not hesitate to adopt the findings of the UK first instance judge, which had followed rigorous testing of the evidence, rather than following the findings of the TBA.
Drawbacks of the current system
It is clear from recent case law that parallel national and EPO proceedings on the same subject matter may result in conflicting findings of facts and, occasionally, in differing interpretations of substantive patent law (where EPO jurisprudence does not disclose a settled approach). This is inherent in a system that allows the national courts to consider validity of a patent after grant and is precisely the same process as found in other jurisdictions with well-developed patent systems. The key difference is that, unlike in the single markets, such as the US and Japan, the absence of a pan-European patent in the European Community (EC) means that patents granted by the EPO for member states of the EC need to be litigated in many different national courts.
European patent dream
There has been talk for many, many years of a unitary patent system for Europe. This would put an end to expensive parallel litigation and divergent decisions on patentability and validity between national courts. After more than 30 years of deliberation and several well-publicised proposals, this dream of complete harmonisation is still some way off.
What the new system will not do, and should not be intended to do, is remove the differences between decisions of the patent-granting body and decisions of the national (or pan-national) courts as final arbiters of the validity of a patent.
However, one thing is clear. Procedural rules will play a key role in any new patent litigation system and could make or break its acceptability to industry. The cases and judicial comment discussed in this article may be said to indicate a degree of scepticism by the English courts towards systems that rely less on the firmly established principles of oral testimony and rigorous cross-examination. No doubt other European courts and judges feel equally attached to their own systems of justice (depending more heavily on written evidence by witnesses and oral submissions by legal representatives alone), which is why the unitary European patent and court is such a daunting project.
What about the EPO?
It is clear from HGS that, in Jacob LJ’s eyes, the UK approach is superior and ought to prevail in any future pan-European patent court. In relation to the EPO, he even went so far as to hint that its more relaxed rules on evidence and the treatment of witnesses may prevent it from accurately determining the facts in issue. He characterised the EPO (somewhat disdainfully) as a ‘coarse filter’ and its proceedings as little more than ‘administrative’.
The UK system certainly provides a ‘finer filter’ at some considerable cost but, perhaps even more importantly, in relation to the EPO, people may need to ‘be careful what they wish for’. As every cook and laboratory technician knows, to add a fine filter can bung up the system completely and that would serve no-one well.