Sofosbuvir - a $2.54bn infringement in the US, but not in Europe

It is a common misconception among inexperienced (typically would-be) patentees that getting a patent for their invention will give them the right to commercialise it. As I have had to explain quite often to new clients, there is a big difference between these two things. A patent only gives you the right to stop others from doing what the claims cover, while being free to put your invention into practice is dependent on not being found to infringe anyone else's patent.

A good recent example of this difference is the verdict from the case of Idenix Pharmaceuticals LLC v Gilead Sciences Inc, as reported here and many places elsewhere. Gilead were found to have infringed US 7,608,597, and have been ordered to pay $2.54 billion in damages. Gilead have their own patent on the infringing drug, known as Sovaldi and Harvoni and with the generic name sofosbuvir, which I have written about here and here in relation to the European patent. Gilead do, of course, have a corresponding US patent for the drug, but this did not help them in their case against Idenix.

What is also interesting about this case though is the contrast that it shows between what can be achieved in the US and what can be done with the same starting material in Europe. Claim 1 of the US patent is very brief, and reads as follows:

1. A method for the treatment of a hepatitis C virus infection, comprising administering an effective amount of a purine or pyrimidine 𝛽-D-2'-methyl-ribofuranosyl nucleoside or a phosphate thereof, or a pharmaceutically acceptable salt or ester thereof.

Claim 1 of the only corresponding granted European patent EP2319856B1, however, is very very long indeed, much too long to be reproduced in full here. It starts off in a similar vein, defining a pharmaceutical composition for use in a method for the treatment or prophylaxis of a Hepatitis C virus infection in a host, but then goes on for a further two pages with various structural features and alternatives. What was possible in the US therefore seems to be impossible in Europe in this case. Not only would sofosbuvir alone not be an infringement of the European patent (I am fairly sure about this, but happy to be proven wrong), but also European courts do not go for such huge damages as are found in the US (although the headline $2.5 billion figure is very likely to be whittled down on appeal).

Another interesting feature of this long-running story is that an infringement action has already been tried by Idenix against Gilead in Europe, but this ran aground when Mr Justice Arnold at the England & Wales High Court found that Idenix's patent EP1523489B1 was invalid (see the IPKat post here), and this was upheld on appeal. So, at the moment at least, it looks like Gilead have done very well in Europe to keep their patent on sofosbuvir and also avoid it infringing anyone else's, but the story is very different in the US. This seems to be largely down to more thorough examination in Europe, both at the EPO and the courts (although this does not always apply: see my recent post here), compared with the generally far more lenient USPTO and more patent friendly court system in the US.

Confusion around G 1/15 & conceptual splitting

Since I reported a couple of weeks ago about the order issued by the Enlarged Board in G 1/15, there have (at the last count) been 69 comments on the post, which for me is quite a lot. There is a similar, though not quite as lengthy, discussion on the IPKat here. It seems that the EB's order has not yet settled the issue around partial priorities, and there remain some questions about how it should be interpreted. We should, of course, get some clarity once the reasoned decision comes out. For now though, there seems to be quite strong disagreement among some commenters about how to assess partial priority.

The disagreement seems to boil down to whether conceptual splitting of a generic OR type claim can result in different priorities for different parts of a generic claim. Some anonymous commenters have said that this cannot be correct, as it cannot be right to assign a priority to a part of a claim that has not been disclosed. Others, however, have said that conceptual splitting is not even needed, since the only requirement is based on what is disclosed in the priority document. My personal view is that the conceptual splitting approach makes more sense, and avoids problems with Article 4C(4) of the Paris Convention, even though it does result in assigning a priority date to a part of a generic claim that is not literally disclosed in a priority document. It is plausible that the EB could come to a different interpretation, although in my view this interpretation is already covered by the words "at least implicitly" in the order.

The issue can probably be best explained by reference to one of the examples in the FICPI memorandum which, as you will recall, was the basis for the decision on partial priorities in G 2/98 and which was totally ignored by decisions from the boards of appeal that went down on the side of literal interpretation of point 6.7 (the "limited number of clearly defined alternatives" bit). The example reads as follows:

b) Broadening of range (temperature, pressure, concentration, etc.).

     Let us assume that a first priority document discloses a temperature range of 15-20° and a second priority document discloses a temperature range of 10-25°. 

     If multiple priorities for one and the same claim are allowed, it will suffice in the European patent application to draw up a claim directed to the temperature range 10-25°. 

     If multiple priorities for one and the same claim are not allowed, the applicant will have to draw up two parallel claims, one directed to the temperature range 15-20° and enjoying the first priority, and another claim directed to the temperature ranges 10-15° or 20-25°, which claim will then enjoy the priority of the second priority document which was the first one to disclose these temperature ranges. 

It should be clear from this example that there is no need to have literal basis for the ranges 10-15° and 20-25° in the second priority document (let's call it P2) to allow priority for these portions to be validly claimed in the later filed European application (let's call it EP). This, to me at least, suggests that the conceptual splitting approach is correct, where priority for the narrow range of 15-20° is from the first priority document (P1) and for the other parts of the range from P2.

I would guess that the reasoning in G 1/15 will need to go over in detail each of the examples in the memorandum (which G 2/98 did not) to establish clearly what they mean by assigning partial priorities. Unless they do, the argument about whether conceptual splitting is allowed or not will continue to rumble on. My prediction, for what it is worth, is that they will go over each of the examples, and find that conceptual splitting is a necessary feature of assigning partial priorities to a generic OR type claim, particularly in cases where priority is claimed from more than one document, such as in the example above. I might, however, be wrong.