30 June 2017

Hit or miss? English court plays Drug Box Jury

It seems that the financial meltdown, which occurred in 2008 and led to austerity measures being imposed worldwide, has reached the pharmaceutical sector.

In the case of Astex Therapeutics Ltd –v- AstraZeneca AB [2017] EWHC 1442 (Ch), Mr. Justice Arnold considered AstraZeneca’s decision to seek to claw back several million US dollars it had paid to Astex Therapeutics (‘Astex’). To be able to decide this meant that he had to consider the terms of a research and collaboration agreement in relation to the search for drugs to inhibit Alzheimer’s disease.

As with many collaborations in the pharmaceutical field, the large pharmaceutical companies have the cash, but less ability to research specific drug applications, whereas the small research companies (often university spin-outs) have the skill and expertise, but not the cash. So, when a dispute like this happens, the commentators tend to reach for the ‘David and Goliath’ metaphor. This may once have applied in this case.  However, Astex was bought by the Japanese group, Otsuka, which has a group turnover of £7bn, so more a case of David, plus his gigantic brother, versus Goliath, who is not looking so big after all.


The Collaboration Agreement was for just over two years, commencing February 2003, after which AstraZeneca could continue researching the fruits of the collaboration by itself (which it did in its’ Swedish centre). That subsequent research identified two compounds (called CD1 and CD2 in this case) as possibilities. AstraZeneca nominated these as ‘Candidate Drugs’ as a result of which a milestone payment was paid to Astex.  Although CD1 failed at the first hurdle, CD2 is currently in the final (Phase III) stage of a clinical trial.  Shortly before the Phase III trials, AstraZeneca wrote to Astex and stated that contrary to its earlier view, neither CD1 nor CD2 were compounds which fell within the definition of the Agreement (Collaboration Compounds), upon which they were obliged to pay further milestone payments.

In essence, despite AstraZeneca’s own conduct in considering that a milestone payment was to be paid, nevertheless, the issue was whether AstraZeneca could get its money back. For those of you with some knowledge of contract law, this is not on the basis of a mistake leading to the rescission of a contract, but rather a question of fact as to whether CD1 and CD2 fell within the definition of ‘Collaboration Compounds’, and there had been enrichment as a result of that mistake. The behaviour and conduct of the parties had no bearing in this particular matter.

In order for the milestone payment to become due, the compound had to be a ‘Hit’ or a ‘Lead’, after which AstraZeneca had to nominate the Collaboration Compound as a CD (and they nominated CD1 and CD2).

Once those milestone payments had been paid, and the compound was on a path to clinical trials, the subsequent payments became eye-wateringly large, from US$1m up to US$5m at the start of the Phase III clinical trial: Hence, why AstraZeneca reconsidered whether any of the compounds actually fell within the definition.

The Collaboration Agreement was 67 pages long, detailed the research steps and specific project concerned, and regarded by the judge as extremely complex.

In 2009, the parties executed an amendment agreement altering some of the terms of the Collaboration Agreement.  The key provision, Clause 2.3, provided that the Collaboration Agreement had expired and all the Hits, Leads and CDs had been agreed by the parties and were attached as a schedule.  That schedule was 127 pages long and contained 1,335 compounds.

‘Can I borrow your CD?’

The trial lasted 13 days, involved 16 witnesses of fact and one expert per side.  The decision is 90 pages long.  In his judgment, the judge skips fairly lightly across the technical background to Alzheimer’s disease and possible inhibitors, but only because he had previously set it out in a judgement of his in 2013, which was itself another 90 page blockbuster. The next large chunk of text delved into the world of drug discovery.

The only part of his judgment that was brief dealt with the law and approach to the construction of the agreement: a now well-trodden path.

The difficulty faced with increasingly lengthy and complex contracts, is that the chances of provisions seemingly contradicting themselves increases. This was apparent in the first part decided by the judge: could the Project have been carried on by AstraZeneca after the Collaboration Term ended?  Although he considered both yes and no answers were possible, the ambiguities of the drafting tipped the balance in favour of AstraZeneca. The significance of the finding of the judge was that it begged the question whether CD2 resulted from the ‘Program’ or arose out of the work AstraZeneca did on their own. The route the compound had to follow, was being an Affinity Hit (‘AFFIT’ –  which was a compound showing promise in binding in the right area, modified during the Program, this being ‘AFFIT Optimisation’) and then being a Hit, which is an AFFIT Improvement having been identified during AFFIT Optimisation, and  selected as a candidate to become a Collaboration Compound.

In effect, the legal definition trail centred around ‘Collaboration Compound’ and which led the judge to conclude that:

  1. Collaboration Compound included all Hits, Lead Compounds, CDs… discovered or identified as a direct result of AFFIT Optimisation….;
  2. An AFFIT is not within the Collaboration Compound definition unless it is selected as a Hit; and
  3. To qualify for AFFIT Optimisation, the chemical structure modification had to have been performed as part of the Program, starting from an AFFIT aimed at identifying Hits.’

The next 40 pages of the judgement (and so the vast bulk of it) looked in minute detail at the collaboration of the parties and the laboratory notebooks of those involved.  Having analysed the actual situation as he saw it, the judge concluded that neither CD1 nor CD2 were as a direct result of AFFIT Optimisation and so did not fall within the definition of Collaboration Compound.

Would you like cash back?

The final part of the judgment was to consider whether AstraZeneca, having accepted by its own conduct that payment was due, and paying it, nevertheless could recover their money. This is wonderfully called ‘a claim to restitution of an enrichment’ that can only be claimed where:

  1. At the time the enrichment was conferred, the payer was mistaken; and
  2. The mistake caused the enrichment to occur and, but for the mistake, the enrichment would not have occurred.

The main evidence, oddly enough, came from AstraZeneca’s own former Vice-President (but who gave evidence on behalf of Astex), in which he reiterated his firm belief that CD1 was a Collaboration Compound. The judge carefully reviewed the process that was followed by those witnesses who made the recommendation that the milestone payment for CD1 should be paid. This turned out to be the former Vice-President himself.  The judge neatly pointed out that the witness was undoubtedly mistaken at the relevant time, not least because he still believed CD1 was a Collaboration Compound when, as a fact, the judge concluded it could not be. So, he was still mistaken, even after judgment.

As for the payment for CD2, it transpires that because of what happened in respect of CD1, the finance department of AstraZeneca merely followed suit because there was no reason to consider that CD2 was to be dealt with any differently to CD1, for which a milestone payment had already been made.

So, AstraZeneca was able to recover the money they had paid.

D. I. V. O. R. C. E.

A song that was certainly a hit for Tammy Wynette in 1968, but which also spells what happens with Collaboration Agreements when money is at stake. And, like many other wealthy divorces that take place in London, God knows how much the legal bill added up to.

What it does highlight, however, is the importance of consistency and clarity in legal documents. Yet even here, in an agreement 67 pages long, there are still errors in drafting, but it is not the error in drafting that ultimately caused the problem: it was the error in legal construction as to whether something did or did not fall within the definition of ‘Collaboration Compound’ and thus whether or not a payment was necessary.

The salutary lesson here is that it is all well and good clever lawyers drafting lengthy, intricate agreements if, in the end, those that must operate under them have no real idea as to how to interpret the drafting.

Whatever else, this particular compound was a miss: at least in the legal sense.