T 1608/13 08-05-2018

I. The patent proprietor has appealed against the Opposition Division's decision, dispatched on 17 May 2013, to revoke European patent No. 1 572 330.

The patent was opposed on the grounds of lack of novelty and inventive step, insufficient disclosure and added subject-matter.

The Opposition Division held that the invention as defined in claim 1 of the patent as granted was not disclosed in a manner sufficiently clear and complete for it to be carried out by a person skilled in the art. Hence, the ground for opposition according to Article 100(b) EPC prejudiced the maintenance of the patent as granted. The Opposition Division did not decide on the other grounds on which the patent was opposed.

II. Notice of appeal was received on 10 July 2013. The appeal fee was paid on the same day. The statement setting out the grounds of appeal was received on 27 September 2013.

III. Oral proceedings took place on 8 May 2018.

The appellant requested that the decision under appeal be set aside and that the patent be maintained as granted or, in the alternative, on the basis of auxiliary request 1 filed with letter dated 26 September 2013.

A request for reimbursement of the appeal fee, made with the statement of grounds, was withdrawn during the oral proceedings.

The respondent requested that the appeal be dismissed.

IV. The following documents are mentioned in the present decision:

D7: European Standard EN 1283, 1996;

D9: "Analysis and Prediction of Sieving Curves for Ultrafiltration Membranes: A Universal Correlation?", A.S. Michaels, Separation Science and Technology, Vol. 15, No. 6, pp 1305-1322, 1980;

D17: "Replacement of Renal Function by Dialysis: A Textbook of Dialysis", Third Edition - Updated and Enlarged, edited by J.F. Maher, Kluwer Academic Publishers, pp 319-321, 1989;

E3: "Basic Principles of Membrane Technology", Second Edition, M. Mulder, Kluwer Academic Publishers, pp 182-189, 2003;

E4: EP-A-0 568 045.

V. Claim 1 of the patent as granted reads as follows:

"A permselective asymmetric hollow fibre membrane for the separation of toxic mediators from blood, comprised of at least one hydrophobic polymer and at least one hydrophilic polymer, wherein a separation layer is present in the innermost layer of the hollow fiber characterized in that said membrane allows passage of molecules having a molecular weight of up to 45 000 Daltons with a sieving coefficient of 0.1-1.0 in presence of whole blood, and has a nominal cut-off of 200,000 Daltons, with a sieving coefficient of 0.1, in water."

VI. The appellant's arguments where relevant to the present decision may be summarised as follows:

In the impugned decision the Opposition Division had held that example 1 in the patent did not exemplify at least one way of carrying out the invention. However, that example provided clear and unambiguous instructions as to all manufacturing steps and parameters needed by the skilled person for manufacturing a membrane according to claim 1. Whether the nominal cut-off in water defined in claim 1 was explicitly disclosed in example 1 was irrelevant, as there was no reason to believe that the membrane of that example would not fulfil all the features of the claim. The respondent had not discharged its burden of proof in that respect.

In the impugned decision the Opposition Division had stated that the patent did not disclose sufficiently how the sieving coefficients in the presence of whole blood could be determined. However, the skilled person knew that D7 formed the basis for determining the sieving coefficients. The Opposition Division had noted that according to D7 both bovine and human plasma could be used to determine them and that the patent did not specify which type of plasma should be used. However, there was no reason to believe that there was a substantial difference between the plasma types cited in D7. Again, the respondent had provided no proof in that respect. The Opposition Division had also noted that D7 prescribed very specific cross-flow and filtration rates in relation to the maximum blood flow rate as defined by the manufacturer. Since the patent was silent on that maximum blood flow rate for the claimed membrane, the sieving coefficients could not be determined. However, the skilled person would always choose maximum blood flow rates between practical operation limits, which were generally known for haemodialysers for medical use, and within those limits the influence of the flow rate on the outcome for a sieving coefficient measured in accordance with D7 was negligible, as also derivable from D17 (page 320, left column, first paragraph). The Opposition Division had also noted that D7 taught how to determine sieving coefficients for a hollow fibre module and not for a single hollow fibre as defined in claim 1. The mass transfer conditions strongly depended on the fibre density, length and flow distribution geometry of the module, which were not specified in the patent. However, the sieving coefficients were a function of the structure of the membrane and were not significantly influenced by the parameters mentioned by the Opposition Division within practical operation limits in the field of haemodialysis. The impugned decision also referred to the fact that the appellant had stated that variations of up to 20% among measurements of sieving coefficient values had to be expected. That was proof that the sieving coefficients as claimed could not be reliably determined. However, such variations could be expected between single measurements, whereas the skilled person knew that several measurements had to be performed in order to reduce standard deviation and improve the reliability of the final (average) measured value. According to the impugned decision, the patent did not disclose how the nominal cut-off in water had to be determined. However, it was common general knowledge (derivable from D9) that dextran would be used for determining the cut-off in a molecular range around 200,000 Dalton, as that was the commonly used tracer having a molecular weight distribution in that range.

In the impugned decision the Opposition Division had stated that the patent did not disclose sufficiently how the invention could be performed over the whole range defined in claim 1. However, the patent did not have to disclose all possible ways to obtain a membrane according to claim 1. A patent claim was intended to provide a generalisation of a specific disclosure, in a good balance between that disclosure and the contribution over the prior art. Rather, the skilled person, knowing standard haemodialysis membranes and starting from the manufacturing conditions and parameters used in example 1 of the patent, would be able to perform slight modifications of those parameters and, with some trial and error, using routine methods, obtain other membranes according to claim 1. Specifically, the given temperature range of 30° to 80° of the polymer solution from which the membrane was to be obtained, as mentioned in paragraph [0041] of the patent, had to be interpreted as a necessary but not sufficient condition in order for a membrane according to claim 1 to be obtained.

The present case was similar to that underlying decision T 339/97, in which a membrane with an exclusion limit for molecules between 30,000 and 40,000 Dalton was claimed. The board in that case had considered that the skilled person was in a position to readily determine the exclusion limit of a membrane, even if somewhat varying measured results were obtained. Such variations might possibly affect the clarity of the subject-matter claimed, but did not render the invention insufficiently disclosed since, in accordance with the findings in decision T 608/07, they would not deprive the skilled person of the promise of the invention. Clarity was not a ground for opposition.

It followed that the invention as defined in claim 1 of the patent as granted was sufficiently disclosed to the skilled person. The case had then to be remitted to the Opposition Division for further prosecution, in order for the remaining issues to be considered by two instances.

VII. The respondent's arguments where relevant to the present decision may be summarised as follows:

The Opposition Division had correctly decided that Article 83 EPC was not fulfilled for two main reasons, i.e. that the patent as granted did not provide any information how the features of claim 1 and example 1 could be determined and that the claimed invention could not be put into practice over its whole scope. According to decision T 585/92, in appeal proceedings it was the appellant who bore the burden of proving that the reasons for the impugned decision were not correct.

According to decisions T 593/09 and T 83/01, if a parameter defining the invention was described in such an ambiguous way as to make its determination or measurement impossible for the skilled person, that resulted in the invention being insufficiently disclosed. The patent did not describe how to measure the claimed nominal cut-off of 200,000 Dalton or the sieving coefficient as defined in the characterising portion of claim 1. It was doubtful that the skilled person would implement the measuring method described in D7. D7 concerned the determination of sieving coefficients of filter modules and not hollow fibre membranes, and used a different definition of the sieving coefficient, according to the equation under point 5.4.2.2. Moreover, E4 disclosed a different measuring method. For any measuring method to be possibly performed it was not disclosed which molecule had to be used, how the membrane had to be arranged in a filter module, and which kind of plasma and which flow conditions had to be used. All of these parameters had an influence on the measurement of the sieving coefficient, as explained in E3, pages 183 and 184. Regarding the dependence of the sieving coefficient on the flow rate, the appellant's arguments based on D17 were to be disregarded, since D17 was not concerned with determining the sieving coefficient in blood or water, but in plasma, which had a different protein content. According to D17 (page 320, last paragraph), the presence of proteins had a strong influence on the sieving coefficient. The appellant's statement that the skilled person would know that dextran had to be used to determine the nominal cut-off as defined in claim 1 was also not credible. The patent did not mention dextran at all, and there were different methods for determining sieving coefficients. Contrary to the appellant's assertions, the sieving coefficient also depended on the geometry of the fibres within a module, as derivable from D17, figure 17 and page 320, first paragraph. The appellant itself, in a letter dated 15 November 2012, had admitted that in the measurement of a sieving coefficient variations of up to 20% had to be expected. Hence, the skilled person could not measure the sieving coefficients defined in claim 1 of the patent as granted with a precision necessary to establish whether a given membrane solved the problem of the invention or, at least, to establish whether the membrane fell within the scope of the claim. This last uncertainty alone, according to the jurisprudence of the boards of appeal (for example T 464/05 or T 256/87), resulted in an insufficient disclosure.

The appellant had not proved that a membrane with the combination of sieving coefficients defined in claim 1 of the patent as granted could be manufactured. In the letter dated 15 November 2012 it had shown that a membrane manufactured according to example 1 of the patent had a sieving coefficient of 0.1 in water already for molecules of 66,000 Dalton. That implied that for molecules of 200,000 Dalton the sieving coefficient would be much smaller, contrary to the requirements of claim 1 of the patent as granted. The patent as granted was also contradictory, in particular concerning the sieving coefficient for molecules of up to 45,000 Dalton. In particular, figure 3b showed that for those molecules the sieving coefficient was 0.8 to 1, contrary to the requirement of claim 1, according to which it had to be 0.1 to 1. Moreover, according to the patent, paragraphs [0043] to [0056], the manufacturing of a membrane depended on so many parameters that it was not feasible, without undue burden, for the skilled person to try to modify the specific manufacturing method proposed in example 1 and work out the invention over the whole scope defined in claim 1. That could also be derived from the membranes obtained according to examples 2 and 3 of the patent, which constituted comparative examples even if they were obtained under almost the same manufacturing conditions as in example 1. In particular it was hardly conceivable that a membrane according to claim 1 of the patent as granted could be obtained in the whole temperature range of 30° to 80° of the polymer solution mentioned in paragraph [0041] of the patent.

Decision T 339/97 was not relevant, in particular because in the present case the membrane was defined by two retention parameters which influenced each other. Moreover, one of these parameters was a single value of the cut-off of 200,000 Dalton, and not a range.

The case should not be remitted to the Opposition Division for further prosecution, but the patent had to be revoked also on the grounds of added subject-matter, lack of novelty and lack of inventive step.