Actavis UK Ltd & Ors v Eli Lilly & Company

2. The issues which we have to decide are: i) The forrn of order for costs of the trial and the appeal; ii) What should happen to the interim payment made by Lilly in the court below, and whether there should be an interim payment of any costs awarded under iii) Whether the order should recite the terms of a letter from Actavis's solicitors dated 17 April 2013; iv) Whether there should be permission to appeal to the Supreme Court. Costs

3. The table below sets out the respective parties' estimates of their total costs of the trial and appeal, and that part of each which is attributable to the Rome Il and the DNI factual issues: Actavis Total costs at trial 3,344,786 Rome Il/DNI factual at (55%) 780,716 (35%) Total costs of appeal 620,388 910,539 Rome Il/DNI factual on appeal 248,000 (40%) 318,689 (35%)

4. The principles on which costs are awarded are well known. First one has to identify the winner. Although Lilly has not won as comprehensively as it would wish to have done, there is no doubt that it is the overall winner. Actavis sought clearance for their products and have failed to obtain the declarations which they sought. The starting point, but no more than that, is that Lilly should recover its costs. by v

5. Next we should consider whether there should be a reduction for any issues on which Lilly has lost, and in respect of any such issues whether there is a reason to go further and order Lilly to pay Actavis' costs of that issue.

6. Lilly has lost on the Rome Il issue, and on the DNI factual issues which flowed from it. These are discrete issues which (as least so far as Rome Il itself is concerned) Lilly initially conceded. The issues were argued by separate leading counsel from those who argued the technical issues. We consider that Lilly should pay the costs of those issues, and not simply have its costs reduced by the amount of those issues, both here and below. On the basis of the figures above, the effect of such an order on the costs of the trial would be to result in a costs order in favour of Actavis, as the sum of the amounts spent by both parties on the Rome IL/DNI factual issues exceeds Lilly's total costs at trial. However, the parties differ widely on the percentage of their costs expended on that issue. We consider that, so far as the trial is concerned, there will have to be an issue based assessment of the Rome Il/DNI factual issues.

7. We also consider that there should be a further reduction of Lilly's costs at trial to reflect the fact that they failed on the direct infringement limb of the patent law issues. Given the limited legal and factual dispute on indirect infringement, it is clear that direct infringement will have been responsible for the lion's share of the costs remaining after the deduction for Rome 11/DNI factual issues is made. Actavis estimate that 14% of Lilly's total trial costs related to indirect infringement. This of course does not reflect the fact that Lilly is also entitled to its general costs of the action. Given that we are directing an issue based assessment of the Rome Il/DNI factual costs, we should express Lilly's percentage recovery in terms of a percentage of Lilly's assessed costs after deduction of the Rome Il/DNI factual costs. We consider that Lilly should recover 25% of those costs, from which Actavis' costs of the Rome Il/DNIfactual issues will also have to be deducted. The balance will be payable to the party to whom it is due. On the figures we have this seems likely to result in a payment in favour of Actavis. We do not consider that this would be an unjust result given that Lilly has lost on large, fact sensitive and expensive issues.

8. We consider that a similar overall regime should apply to the appeal. Here, however, the parties are less at odds over the percentage expended on the DNI/factual issues. There is no need to direct an assessment of the costs of the Rome Il/DNI factual issues in the appeal. Accepting Lilly's estimate, 35% of their costs fall to be deducted for this issue. Actavis estimate that about 25% of the total costs of the appeal were spent on indirect infringement If we were to take this figure, Lilly would recover 25% of E910,539, or E227,634. From this we would have to deduct 248,000 being Actavis' costs of the Rome Il/DNI factual issues. Given that Lilly was the overall winner, we consider that overall justice would be done if there were no order as to costs of the appeal. Interim payment

9. By his order dated 3 June 2014, Arnold J ordered Lilly to pay the sum of E 1,839,632.30 on account of the costs of the trial. This interim payment was approaching 50% of Actavis' total costs of E3.34 million. Although the overall outcome of our adjusted costs order of the trial is not yet clear, it is clear that the interim payment may and probably will exceed Actavis' likely recovery. by 1'

10. The effect of the orders we have made based on the figures in the table above would be a payment to Actavis of about El .45 million. The figures may look very different after the assessment which we have directed. Interim payrnents should be cautious, and particularly so in this case where there is yeat uncertainty as to the final outcome on costs. We accordingly vary the order for interim payment below to substitute the sum to E 350,000. We order repayment of such sum as will reduce the amount paid to that figure. Recital l l . Lilly's draft order inserts a recital of a contractual undertaking. We do not consider any purpose is served by this recital, and will not include it in our order. Permission to appeal

12. Actavis point to two issues on which they submit they should have perrnission to appeal. The first is the relevance of the prosecution history to an issue of claim construction. The second raises a point about the relationship between direct and indirect infringement. We do not consider that the first point is relevant to the basis on which we have found in favour of Lilly. On the second point it is argued that it is anomalous that something which does not fall within the claim on direct infringement (because it cannot be predicted that it will work) nevertheless amounts to an indirect infringement if it does. That way of putting the point obscures the fact that, on the indirect infringement case, no question of prediction of the behaviour of a variant arises. It is the claimed material itself which is made, not a variant of it. The proposed appeal on this point does not therefore raise an arguable point of law. Draft order

13. The parties should now be in a position to produce a draft order for submission within 7 days to reflect the main judgrnent and the matters resolved by this supplementary one.