Skip to content

NL – Astrazeneca v. Sandoz / Fulvestrant / Appeal

05 Dec 2018

Astrazeneca AB v. Sandoz B.V., Court of Appeal The Hague, The Netherlands, 27 November 2018, Case no. 200.237.828/01, with thanks to Willem Hoyng and Theo Blomme, HOYNG ROKH MONEGIER, for sending in the case

The Court of Appeal of The Hague has upheld AstraZeneca’s patents EP 1 250 138 B2 and EP 2 255 573 and ruled that Sandoz had infringed these patents by marketing its generic Fulvestrant Sandoz 50 mg/ml solution for injection in a prefilled syringe. In its decision of 27 November, the Appeals Court quashed the decision of the District Court (which can be read here) which had invalidated AstraZeneca’s patents for lack of inventive step.

The decision focuses on a Swiss type claim, present in both patents, covering the use of fulvestrant for the preparation of a pharmaceutical formulation for intramuscular injection for the treatment of breast cancer. The claimed formulation has a total volume of 5 – 5.25 ml, a total amount of fulvestrant of 250 mg or more and a pharmaceutically acceptable alcohol consisting of 10 weight% ethanol, 10 weight% benzylalcohol, as well as 15 weight% benzylbenzoate. Furthermore, the claimed formulation is contained in castor oil and it is adapted to achieve a therapeutically significant plasma concentration of fulvestrant during a period of at least two weeks.

The Court of Appeal follows the District Court’s reasoning as to the closest prior art and the formulation of the objective technical problem. The closest prior art is an article by Howell, which discloses that a formulation of 250 mg fulvestrant in 5 ml castor oil has proven efficacious in Phase II clinical trials in the treatment of breast cancer, that one injection achieved a sufficient plasma concentration of fulvestrant during one month, and that it was well tolerated. Howell does not disclose the cosolvents and their relative concentration.

The objective technical problem to be solved by the skilled person was therefore to obtain a formulation in which 250mg fulvestrant is dissolved in 5 ml castor oil which is suitable for treating breast cancer, does not precipitate and provides extended release to the extent that a therapeutically significant plasma concentration is achieved during at least two weeks after intramuscular injection.

Similar to the District Court, the Court of Appeal finds that the skilled person would, in the search to solve the problem, take into account McLeskey, which discloses the claimed fulvestrant formulation. However, unlike the District Court, the Appeals Court ruled that McLeskey does not teach the use the formulation for the treatment of breast cancer, since McLeskey merely discloses the use in mice, and in a different context, namely to test whether estrogen affects the growth of tumor cells when the mice’ ovaries have been removed.

Sandoz has not sufficiently substantiated that results found in mice have predictive value for the release and precipitation profile in humans. Interestingly, the Court did find, based on an statement by AstraZeneca’s expert Prof Robertson, that results of tests in rabbits, which were mentioned in the description of EP138, were able to predict the effect in humans.

The fact that Riffkin, another publication, disclosed that two of the co-solvents from McLeskey were suitable to dissolve a parentally administrated steroid such as fulvestrant, and that the third McLeskey co-solvent was a commonly used solvent in formulations for human use, did not teach the skilled person that the use of such formulation in humans would have the required effect, i.e. lead to extended release and prevention of precipitation.

Furthermore, the fact that the formulation in McLeskey contained 20% alcohol, which is volatile, rendered it likely that much of the alcohol would diffuse from the injected solution, leaving an oversaturated solution of fulvestrant in castor oil and that, as a result, fulvestrant was likely to precipitate.  According to the Appeals Court, McLeskey therefore even contained a contraindication to consider the McLeskey formulation as a solution to the objective technical problem.

The Court of Appeal notes that both the undisclosed co-solvent formulation used in Howell on the one hand, which Howell shows achieves the required effects in humans, and the McLeskey formulation on the other hand originated from (the legal predecessor of) AstraZeneca. Although this may create a presumption that the same formulation was used, it did not render the claimed invention obvious, since the aforementioned contraindication in McLeskey would have steered the skilled person away from this presumption.

Finally, the Court of Appeal also rejected Sandoz’ argument that the skilled person was in a ‘try and see’ situation. According to EPO Case Law, this requires that the skilled person already clearly envisages a certain compound or group of compounds, which would then routinely be tested. The Court of Appeal found that in the present case, there was no incentive (but rather a contraindication) to use the McLeskey formulation, so it cannot be established that the skilled person had already ‘clearly envisaged’ this formulation. The Court of Appeal therefore ruled that the subject matter of the patents is non-obvious. It also rejects Sandoz’ arguments regarding lack of novelty in light of McLeskey, since McLeskey does not disclose an effective treatment of breast cancer or humans, not does it disclose intramuscular injection.

Sandoz’ invalidity arguments fail. The Court of Appeal confirms the validity of AstraZeneca’s patent and quashes the first instance decision. The injunction which was awarded in first instance is reinstated.

The judgment (in Dutch) can be read here.

Headnote: Jaap Bremer, BarentsKrans