T 0703/18 (Infant formulas containing DHA and lutein/ABBOTT) 23-02-2023

Reasons for the Decision

1. Patent

1.1 The patent relates to an infant formula containing combinations of lutein and docosahexaenoic acid. The composition is stated to promote retinal health and vision development in infants (paragraph [0001]).

1.2 Paragraph [0007] discloses that infant formulas typically contain less than 20 myg/liter of lutein, most of which is inherently added with fats and oils. According to the patent, the invention resides in recognising that bioavailability of lutein from infant formula is low. The lutein concentration in an infant formula has been higher than that found in human milk to achieve the same plasma lutein concentration as in breast-fed infants.

2. Article 100(a) EPC - Inventive step

2.1 In the decision under review, the opposition division concluded that claim 1 involved an inventive step. The reasoning was as follows.

- D16 was the closest prior art. The distinguishing feature of claim 1 over example 1 of D16 was the higher concentration of lutein.

- None of the cited documents referred to bioavailability of lutein but rather to that of trace elements and lipids. The problem was to provide an infant formula that led to the same plasma lutein concentration as in breast-fed infants. In such a case, "identifying the problem in itself leads to an inventive step (problem invention)" (decision, page 6).

- The prior art did not teach the problem identified. There would have been no motivation to increase the amount of lutein.

2.2 The appellant argued that the opposition division's formulation of the problem and finding on obviousness was incorrect.

2.3 Closest prior art D16

2.3.1 The parties agreed to use D16 as the closest prior art.

2.3.2 D16 discloses infant formulas containing lutein and zeaxanthin. Paragraph [0016] reports a study in which the total lutein and zeaxanthin content of human milk samples from 450 women in nine countries was analysed. The individual concentration of the two carotenoids was between 6 myg/liter and 230 myg/liter. Dependent claim 2 is directed to an infant formula composition comprising these concentrations of lutein and zeaxanthin.

2.3.3 Example 1 of D16 discloses a nutritional infant formula having a concentration of lutein and zeaxanthin of 25 myg/liter. The concentration of docosahexaenoic acid is calculated to be 75.6 mg/liter.

2.3.4 The parties agree that example 1 of D16 does not disclose the lutein concentration set out in claim 1 (i.e. 75 myg/liter).

2.3.5 As correctly pointed out by the respondent, a direct consequence of the low concentration of lutein in example 1 is that the weight ratio of lutein (myg) to docosahexaenoic acid (mg) of claim 1 is also not disclosed in D16.

2.4 The problem to be solved proved to be a contentious issue.

2.4.1 The appellant argued that the technical problem was to provide an infant formula suitable for newborn infants.

2.4.2 For the respondent, however, the correct technical problem was the one identified in the decision under appeal. In its view, the patent involved one of the rare cases of a "problem invention". It had not been recognised in the art that bioavailability of lutein from formula milk was lower than that achieved by human milk. Once this was known, the solution would have been obvious.

2.4.3 As the respondent acknowledged, "problem inventions" are rare. One reason for this may be that they are somewhat at odds with the problem-solution approach. It is generally accepted that the formulation of the technical problem should not contain pointers to the solution or partially anticipate the solution. In contrast to this, "problem inventions" tend to do both.

2.4.4 A relevant decision dealing with "problem inventions" is T 2/83. In point 6 of the Reasons, the following is stated.

"The discovery of a yet unrecognised problem may, in certain circumstances, give rise to patentable subject- matter in spite of the fact that the claimed solution is retrospectively trivial and in itself obvious ('problem inventions') ... It appears however, that whenever the modification of a known device involves no real choice in the direction of a clearly desired improvement, i.e the skilled man is in an inevitable 'one-way-street' situation, the additional provision of a yet unsuspected 'bonus' or side effect, which may be interpreted as a solution of a yet unknown problem, should not necessarily be decisive for patentability."

2.4.5 Although this passage of T 2/83 refers to a device, it is not apparent why the reasoning in it would not apply also to claims directed to compositions.

2.4.6 The decision T 2/83 concedes that the discovery of an unrecognised problem may in certain circumstances give rise to patentable subject-matter. This may be so even though once the formulation of the problem is accepted, the question of whether the solution was obvious becomes irrelevant. A situation may arise in which, if a subject-matter claimed is assessed as a "problem invention", an attack based on lack of inventive step can be successfully directed only against the recognition of the problem, not against the claimed solution. At the same time, T 2/83 makes it clear that in the context of a clearly desired improvement, side effects which may be interpreted as a solution of a yet unknown problem shall not be decisive for patentability.

2.4.7 In the case in hand, the following has to be taken into account when formulating the technical problem.

2.4.8 First, while the issue with bioavailability of lutein is new information, bioavailability of trace elements and lipids from infant milk has been well investigated. This was thoroughly discussed in the oral proceedings before the opposition division (minutes, points 7.1 and 7.2) and in the decision under appeal (e.g. pages 5 and 6).

2.4.9 Second, the "gold standard" for preparing an infant formula is and remains human milk. Therefore, when formulating the problem to be solved, one must draw closely on the teaching available in the art on human milk. D16 itself discloses that the individual concentration of lutein and zeaxanthin in human milk is distributed over a wide range (6 myg/liter to 230 myg/liter) and teaches the use of these concentrations in an infant formula (claim 2).

2.4.10 Third, the patent in suit relates to preparing an infant formula. The target group for the formula is not restricted in claim 1, but preterm infants are explicitly mentioned throughout the patent. Moreover, the patent also makes reference to newborns, i.e. infants of 0 months of age (paragraph [0013]). In other words, the formula of the patent in suit has to be suitable for newborns and infants born prematurely.

2.4.11 Considering all this, it is not justified to accept the formulation of a "problem invention", as suggested by the respondent. Instead, the technical problem has to be regarded as that of providing a nutritional formula (with lutein) suitable for infants, including newborns.

2.5 Obviousness

2.5.1 The skilled person tasked with solving the problem would turn to D18. This document relates to a study in which lutein and zeaxanthin concentrations in commercial infant formulas and human milk samples are investigated. Lutein and zeaxanthin are described as important protective factors in retinal pigment epithelium of newborn infants (page 91, left column).

2.5.2 Among the results discussed in D18, the following are highlighted.

- The milk of mothers of newborn infants contains much higher levels of lutein than those found after several days of lactation in the milk of the same mother (figures 1 and 2).

- The median concentration of lutein found in expressed human milk is calculated as 109 myg/liter. This value is 8 to 9 times greater than the concentration of zeaxanthin, which is calculated as 12.5 myg/liter (table 1). These values are obtained over a period longer than just the first days of lactation. In other words, the relatively high concentration of lutein is found not only in the first days of lactation.

- The highest lutein concentrations found in commercial infant formula (e.g. calculated as 205.3 and 157.7 myg/liter) are found in milks formulated for preterm infants. They potentially require a higher concentration for protection against oxidative damage to the eye (page 96, right column).

2.5.3 In short, D18 shows that the concentration of lutein found in human milk, and especially in milk of mothers of newborn infants, is considerably higher than that disclosed by example 1 of D16. In view of this, the solution that the skilled person would have provided is to increase the concentration of lutein. This has the inevitable side effect that the bioavailability of lutein is increased. Increasing the amount of lutein in an infant formula is a straightforward exercise. D16 suggests the addition of a commercially available ingredient (paragraph [0014]).

2.5.4 The appellant argued that the skilled person would have chosen a concentration of 75 myg/liter or higher. The board has no reason to doubt that this is a value that the skilled person would consider. This value is within the range suggested in claim 2 of D16.

2.5.5 Be that as it may, irrespective of whether the amount of lutein selected is 75 myg/liter, 109 myg/liter (as discussed in D18, see point 2.5.2 above) or even as high as 230 myg/liter (the highest amount suggested in D16), with any of these concentrations the weight ratio of lutein (myg) to docosahexaenoic acid (mg) called for in claim 1 would be fulfilled.

2.6 Therefore, claim 1 as granted lacks inventive step. The ground for opposition under Article 100(a) and 56 EPC is prejudicial to maintenance of the patent.