T 1021/08 (Human albumin/GRIFOLS SA.) 01-12-2011

Reasons for the Decision

Written submissions

1. In the present case both the appellant and respondent II filed additional written submissions after, respectively, filing their statement of grounds of appeal and replies (see paragraph IV above). In the appellant's case, there were two additional submissions. The only written submissions which are necessarily taken into account are those referred to in Article 12(1) RPBA (to the extent they are relevant and comply with Article 12(2) RPBA - see Article 12(4) RPBA), namely an appellant's notice and statement of grounds of appeal and the respondents' replies, each of which should contain a party's complete case (see Articles 12(1)(a)(b) and 12(2) RPBA). Any other submissions, unless answering a communication from the board (see Article 12(1)(c) RPBA and again subject to Article 12(4) RPBA), are amendments to a party's case and admissible only at the board's discretion (see Article 13(1) RPBA). No explanation was given for filing any of the late submissions and none of the parties concerned made any request for leave to amend its case thereby.

2. The cited provisions of the RPBA quite clearly foresee only one written submission from each party supplemented as necessary by answers to communications (if any) from the board. They do not foresee, and there is no right to, responses to the reply or any further exchanges of written submissions. Indeed the current RPBA were intended inter alia to prevent such "ping pong" submissions (see the full summary of the legislative history in T 1621/09 of 22 September 2011, points 25 to 34). While the board was able to decide the present case without needing to decide on their admissibility, such additional written submissions beyond those envisaged by the RPBA add to the work of the board (and the parties), tend to delay appeal proceedings, and could be the subject of costs orders under Article 16(1) RPBA.

Main request

Sufficiency of disclosure (Article 83 EPC)

3. It stated in the patent in suit that citrate is the agent responsible for the increase in aluminium content over time in therapeutic albumin solutions stored in glass, and that the removal of this citrate down to a specific level below 0.5 mM, preferably less than 0.037 mM, the latter being the detection limit (see page 4, lines 39-40 of the patent) is desirable for lowering aluminium release from the glass container upon storage.

4. In order to carry out the process according to claim 1 and put into practice the medical use according to claim 2, the skilled person has to lower the citrate concentration in the therapeutic human albumin solution to the desired level.

5. Therefore, the question arises whether or not the technical information provided by the patent in suit, supplemented by common general knowledge, enables the average skilled person to arrive at levels of citrate below 0.5 mM, preferably less than 0.037 mM, without undue burden of experimentation.

6. As regards the technical information provided by the patent in suit, the respondents argue that the skilled person could not derive from the patent any causal relationship between the lowering of the citrate concentration and diafiltration, whereas the appellant denies this contention by relying on the passages on page 2, lines 55-57 and page 3, lines 25-30 and on Example 1 (in particular on page 3, lines 54 and 58 and page 4, line 4 ("citrate concentrations obtained") and line 7 ("final citrate concentration")) of the patent.

7. It is stated in the patent in suit on page 2, lines 55-57 that diafiltration removes excess salts, aluminum and ethanol, including "low-molecular-weight compounds" and that citrate is always present in Cohn fraction V used as the starting material in the albumin preparation method (see page 3, lines 25-30 of the patent in suit). Thus, in the board's view, these two passages cited by the appellant merely confirm that diafiltration was useful to remove the "low-molecular-weight compounds", without pointing to diafiltration as the method of choice to reduce the citrate concentration to trace levels. Diafiltration is also referred to in paragraphs [0013], [0015] and [0023] of the patent, however, this is done in the context of the reduction of the aluminium and ethanol concentrations. It is also stated in paragraph [0023] that despite the the solution of the starting albumin fraction V had been diafiltered to eliminate the aluminium to the extent of values of less than 200 ppb and even 50 ppb or less, a citrate contamination exceeding the value of 1 mmol/liter (1 mM) persisted. This passage of the patent would thus rather lead the skilled person to assume that diafiltration was not suitable for further reducing citrate concentration below 1 mM.

8. The appellant also relies on Example 1 of the patent, in particular on page 3, lines 54 and 58 ("[A]fter diafiltration ....the concentration of each batch was determined") and on page 4, lines 4 and 7 ("[T]he citrate concentrations obtained" and "final citrate concentration") for arguing that the patent made it clear that diafiltration was the only step responsible for lowering the citrate concentration.

9. However, while the cited passages of Example 1 mention diafiltration, this is done in combination with other steps, such as the "adjustment to 20% protein concentration" (see page 3, line 54 of the patent).

10. The board is of the opinion that the skilled person could reasonably assume that some additional steps, such as e.g. the addition of water for adjusting the protein concentration to 20%, might have altered the citrate concentration. The appellant argues that the skilled person would have performed such adjustment such that only the albumin concentration changed, while all the other solutes remained constant. In the board's view, however, keeping the concentration of all the other solutes constant would have necessarily implied adding a solution containing citrate, when the scope of the invention was its removal, which would not make sense.

11. A further line of argument of the appellant was that the skilled person would have considered that the three batches having different citrate concentrations on the left of Table 1 of the patent (0.71 mM, 0.12 mM and 0.037 mM) were the result of diafiltration with increasing length of time, with the value "0.71 mM" representing too short a diafiltration.

12. However, it cannot be derived from the patent -either explicitly or implicitly - that the three samples were treated differently. Rather, the fact that measurement of the citrate concentration took place after 14 days and 10 hours (see page 3, line 57 of the patent) does not suggest a diafiltration where the citrate concentration was measured at different times for each sample.

13. Moreover, the board observes that Example 1 fails to provide any information as to the citrate concentrations prior to diafiltration. The appellant admitted in paragraph 4.2 of its Grounds of Appeal that the amount of citrate initially present will be different from sample to sample. Without this information the skilled person could not deduce that a reduction of the citrate concentration had taken place. Rather, he/she was taught by Table 1 of the patent that the same process described in Example 1 apparently yielded three batches having different citrate concentrations, one of which (0.71 mM) did not reach the claimed upper limit of 0.5 mM. The reasons for this failure were not evident to the skilled person. In view of this further uncertainty, the skilled person could not unambiguously conclude that the diafiltration step in Example 1 was responsible for the decrease, if any, in citrate concentration. The skilled person could also not exclude that techniques for citrate reduction alternative to diafiltration had been used, such as chromatography or washing the albumin solution with water and concentrating on an Amicon® or Millipore® cut-off filter.

14. In conclusion, it is the board's view that no causal relationship can be derived from the disclosure in the patent in suit between the lowering of the citrate concentration and the diafiltration step mentioned on page 3, line 53. As a consequence, the skilled person was not in a position to reproduce any of Examples 1 to 3 of the patent in suit, the more so as no indication was given under which conditions (pH, buffers, volumes, time, cut-off, etc.) diafiltration had to be carried out.

15. As regards performing the claimed process on the basis of the patent supplemented by common general knowledge, the appellant argues that once the size of the component to be removed, its maximum end concentration and the diafiltration technique as the method of choice were identified, putting the claimed subject-matter into practice was a matter of routine trial and error. In other words, in the appellant's opinion, the skilled person would modify the diafiltration step referred to in Example 1 of the patent (see page 3, line 53 et seq.) by subjecting a solution of albumin intended for intravenous administration to diafiltration against a solution isotonic with human blood (corresponding to 150 mM NaCl), in order to avoid any change in the ionic concentration of the solution, and automatically arrive at levels of citrate below 0.037 mM.

16. In support of this contention, the appellant refers to the BPL product label in Annex MB-2 of document D4, which shows that the final therapeutic albumin solution comprises 145 mmol/l sodium (see the BPL product label) and document D43 (see under "CLB Product Division/Finished Product Specification No. SE H2200-101/5, point 7.9.1: "sodium 130-160 mmol/l"). The appellant also points out that it was known from document D6 (see page 9, lines 17-18) that in order to obtain reduced aluminium in an albumin preparation, diafiltration against a saline solution was the method of choice.

17. However, in the board's judgement, only the final therapeutic product needs to be isotonic. This does not necessarily imply that the albumin-containing solution needs to be kept isotonic throughout the whole purification process. In fact, there is no suggestion in the patent in dispute, nor in the prior art, that the albumin-containing sample should be kept isotonic throughout the whole process. Rather, page 3, line 40 and page 4, line 35 of the patent ("the albumin solution is adjusted to be stable and isotonic" and "final albumin adjusted by...and isotony (sodium chloride)") suggest that NaCl should be added at the end of the whole albumin preparation process. If the solutions had been already isotonic, as the appellant maintains, there would be no need for them to be "adjusted to be ...isotonic". Moreover, document D2 (see the first paragraph under the heading "The ion exchange chromatographic process") shows that pyrogen free water (PFW) was a perfectly suitable diafiltration solution.

18. Assuming, for the sake of hypothesis, that paragraph [0023] of the patent would have suggested to the skilled person that the diafiltration processes for removing the aluminium of the prior art were too short for removing citrate (the concentration of which remained > 1mM), he/she would have tried to perform diafiltration more exhaustively. However, in view of the "final" adjustment of isotony highlighted in point 17 supra and the absence in Example 1 of any hint about the addition of NaCl during diafiltration, the skilled person could have reasonably used water without any addition of salt as the diafiltration buffer. But no reduction of the citrate concentration below 0.037 mM would have taken place, as shown by Exhibit A filed with the respondent 2's submission dated 18 March 2005. The latter is an experimental report about a comparison of citrate reduction in albumin under different diafiltration conditions. Experiment 1 of Exhibit A, involving only water as the diafiltration buffer failed to reduce the concentration of citrate below 0.99 mM (see "Results").

19. Finally, the appellant argues that paragraph [0025] of the patent explicitly teaches the skilled person that citrate associated strongly to albumin and that this association could be disrupted in the presence of an ionic charge (hence the necessity of using saline solutions), liberating the citrate for removal by an appropriate method such as diafiltration.

20. However, in the board's opinion, paragraph [0025] of the patent does not address the problem of reducing the concentration of citrate but rather provides a tentative explanation (see the expressions ""not well known at present" and "is assumed") of the possible mechanism underlying the aluminium release from the glass, postulating the existence of an albumin-citrate complex, which complex could be dissociated by an "ionic charge". Furthermore, the fact that citrate associated strongly to albumin was known to the inventors of the patent in suit only (who used 0.25 M NaCl to dissociate it; see point 23 infra), not to the skilled person reading the patent.

21. Therefore, in the board's view, the skilled person would have taken paragraph [0025] with caution, or at most as an invitation to investigate experimentally how (and how strongly) citrate interacted with albumin. As admitted by the appellant during the oral proceedings, elucidating the nature of the albumin-citrate complex (electrostatic, lipophilic, etc.) was a prerequisite to design a method for its removal. It was critical for the skilled person to establish how it could be dissociated and/or which counter-anion could displace citrate from the complex, given that all this information was missing in the patent in suit.

22. It is true that the prior art referred to an albumin-aluminium complex, from which aluminium could be displaced by means of a monovalent metal ion (see Example 1 of document D6, read in combination with page 4, lines 3-6 of the same document, and paragraph [0014] of the contested patent, referring to "ionic displacement" in the prior art techniques. But even if the skilled person had combined the teaching of document D6 with the suggestion made in paragraph [0025] of the patent, difficulties could have arisen because citrate was not a multivalent metal ion but, depending on the pH, a multivalent anion of an organic tricarboxylic acid. Moreover, selecting the "ionic" diafiltration route implied that the skilled person had to proceed against the teaching provided on page 3, line 40 ("adjusted to be ...isotonic") and page 4, line 35 ("adjusted by... isotony (sodium chloride)" that ions had to be added only at the end of the whole albumin preparation process.

23. Finally, even if the skilled person used an isotonic solution (150 mM NaCl) as a diafiltration buffer, as the appellant maintains, doubts arise that he/she could reach the citrate concentration limit of 0.037 mM within reasonable time lags, especially in the case of highly concentrated albumin solutions. This is because there is evidence before the board (see the paragraph bridging pages 4 and 5 of the submission dated 6 November 2000 by the then applicant) that the inventors of the patent in suit in reality made use of a 250 mM NaCl solution as a diafiltration buffer. This critical technical information is missing in the patent in dispute.

24. In conclusion, the patent in suit provides no details which could facilitate the repetition of the work. On the contrary, owing to its numerous omissions, the skilled person relying on the information given in the patent, whether supplemented by common general knowledge or not, would not have achieved the desired final goal. Hence, the only alternative left was to ascertain the information missing in the patent by a thorough investigation -undue for an average skilled person- aimed at finding out the parameters (pH, which ions, ionic force, run time, albumin concentration) under which it could be possible to reduce citrate to trace levels. Thus, the requirements of Article 83 EPC are not satisfied.

Admissibility of the appellant's auxiliary request

25. The main request having been found not to meet the requirements of Article 83 EPC, the board considered the admissibility of the auxiliary request filed on 28 October 2011. The appellant submitted in its letter of that date that the auxiliary request dealt with objections of the respondents under Article 123(2) EPC and indeed the request appears to have been limited to addressing those objections. Thus the respondents' submission that it did not address the issue of sufficiency of disclosure under Article 83 EPC appears correct. Therefore the auxiliary request would not have assisted the appellant and, apart from any other considerations such as the filing of the request at a very late point in the appeal proceedings and the fact that it could have been but was not filed in the first instance proceedings, no purpose would have been served by admitting it into the proceedings.

26. The only substantive issues dealt with in the decision under appeal were those of Article 123(2) EPC and 83 EPC. In view of the negative conclusions arrived at by the board in respect of sufficiency of disclosure (see point 24 supra), no need arises for the board to deal with Article 123(2) EPC.