T 1876/21 (Use of a nutritional composition having a low protein amount/NESTLÉ) 25-10-2023

Reasons for the Decision

1. Novelty (Article 54 EPC) - main request

1.1 Independent claim 1 is drafted in the format of a second medical use in accordance with Article 54(5) EPC.

1.2 It is common ground that document D7 discloses nutritional compositions (for infants) having the same composition as recited in claim 1. It is also undisputed that the claimed therapeutic (preventive) effect is disclosed in document D7. The nutritional compositions in claim 1 and D7 contain a reduced protein content compared with "standard" infant formulas. The purported effect is the prevention of obesity and related medical conditions later in the life of the infant fed with such compositions in the first year of its life. Hence, there was no dispute between the parties that the remaining features of claim 1, besides the treated group of subjects, are already disclosed in D7. The disclosure of those features in D7 is correctly described in section 4.4.2 of the decision under appeal (see claims 1 to 3 and 13 to 15 and page 4, lines 8 to 9 of D7).

1.3 The only point of dispute between the parties was whether the patient/subject population as specified in claim 1 constituted a new group of treated subjects in the context of an otherwise known second medical use of a compound or composition (Article 54(5) EPC).

1.4 Interpretation of claim 1

The board agrees with the opposition division's assessment that, when interpreting the claim in the light of the description, the reduction of the risk referred to in claim 1 has to be assessed in comparison with a subject population fed with a corresponding infant formula with a higher protein content (see paragraphs [0006], [0114] and [0115] of the patent; cf. point 3.4 of the decision under appeal).

Likewise, it follows from paragraphs [0036] to [0039] of the patent that the terms "non-obese mother" and "non-overweight mother" refer to mothers having a body mass index (BMI) of below 30 or below 25, respectively, prior to establishment of pregnancy. Consequently, the following assessment of novelty will be based on this interpretation.

1.5 According to established case law of the boards of appeal, the use of the same compound in the treatment of the same disease for a particular group of subjects, can nevertheless constitute a novel therapeutic application, provided that

i) it is carried out on a new group of subjects which is distinguished from the former by its physiological or pathological status (see Case Law of the Boards of Appeal, tenth edition, 2022, I.C.7.2.4.b; cf. T 19/86, OJ 1989, 24; T 233/96).

In decision T 233/96, as a further requirement for establishing such an aforementioned novel therapeutic application, it was held that ii) the establishment of the novel group must not be arbitrary. This requires that there is a functional relationship between said status and the therapeutic effect obtained (see e.g. T 233/96, point 8.7 of the Reasons; cf. T 694/16, point 5.20 of the Reasons). The opposition division based its decision regarding novelty on said criteria i) and ii) (see point 4.4.3 of the decision under appeal).

1.6 With regard to the first point to be elucidated, point i), the board makes the following observations.

1.6.1 D7 is directed to nutritional compositions for infants at risk of developing obesity later in life. The compositions comprise a protein source, a lipid source and a carbohydrate source and have a protein content of less than 1.8 g/100 kcal and an energy density of less than 650 kcal/litre (see e.g. claim 1).

1.6.2 At the same time, D7 already discloses two groups of subjects expressly, namely infants born to obese mothers and infants born to overweight mothers (see the paragraph bridging pages 2 and 3). In this document, it is also mentioned that for these groups it may be possible to reduce the risk of future obesity by feeding the at-risk infant with the nutritional compositions. The remaining third group, i.e. infants born to mothers who are not overweight and not obese at establishment of pregnancy ("lean mothers"), is also mentioned in D7 on page 1, lines 25 to 27.

1.6.3 D7 states that the administration of the formulas is particularly (and hence not exclusively) proposed for infants at risk of developing obesity later in life. The following is stated in the corresponding passage of D7 on page 3, last paragraph: "Without wishing to be bound by theory, the inventors believe that for infants at risk of developing obesity in particular, feeding a nutritional composition with a controlled protein and energy content which is moreover preferably relatively rich in the satiety-inducing protein casein could counteract any tendency on the part of the infant to overfeed, particularly as regards protein intake, whilst supplying sufficient quantities of nutrients essential for growth and development and resulting in a growth pattern similar to that observed in breast fed infants" (emphasis added by the board).

1.6.4 Whilst the aforementioned third group of subjects (mothers' BMI < 25) is mentioned in the introductory part of D7, it seems that D7 does not directly and unambiguously disclose that the nutritional compositions in D7 are administered to said third group of subjects, i.e. infants born to non-overweight and non-obese mothers.

1.6.5 However, the board concurs with respondent 2 that neither the patent nor documents D12 and D13 corroborate that the groups of infants born to non-obese and non-overweight mothers (i.e. having a BMI < 25 prior to establishment of pregnancy) exhibit a physiological status that could be discerned from that of an infant born to an overweight mother, which is unambiguously disclosed in D7 as an infant at risk of developing obesity later in life.

1.6.6 This lack of distinction between the physiological status of the known treatment group "infants born to overweight mothers" and the group recited in claim 1 is aggravated by the fact that claim 1 encompasses children born to mothers having developed gestational diabetes mellitus (GDM) and/or gestational overweight during pregnancy. These conditions are further factors that influence the pathophysiological/endocrinological status and in particular fat mass in infants. This is derivable from the first full paragraph, left-hand column, on page 1101 of D12 and section 2.1 of decision T 321/15 (D17).

1.6.7 With regard to the latter point, the appellant argued that GDM was a rare event. GDM only occurred in 3 out of 1000 lean mothers during pregnancy. The appellant stated in the oral proceedings before the board that this had been reported in a clinical study in 2010. GDM thus had no bearing on the conclusion that it could be said with high certainty whether an infant had been born to a non-obese and non-overweight mother.

However, no evidence of this assertion of fact has been presented. Even assuming that this argument were correct, it would not be persuasive. It cannot undermine the conclusion that more than just a few individuals are exposed to other risk factors for developing increased fat mass in the first month of life in utero besides the mother's BMI at conception.

1.6.8 The appellant also stated that it was clear that the obese mothers in D12 and D13 did not make up 90% of mothers having a BMI >= 25 at establishment of pregnancy. This was not the way in which clinical studies were designed. This expectation is reasonable; however, it cannot invalidate the conclusion that the larger group of infants born to mothers having a BMI >= 25 at conception already greatly overlaps with the subject group in claim 1 in terms of their relative fat mass.

This is clear from the data presented in Table II in D12 and in Table 2 in D13. It thus has to be expected that the corresponding distribution curves obtained for the infant groups as characterised by "mother's BMI < 25" vs. "mother's BMI >= 25 < 30" would converge even further. This would mean excising the group of the infants born to obese mothers, having a BMI of at least 30. No data for the group of infants born to overweight mothers only are provided in D12 or D13. Therefore, D12 and D13 fail to demonstrate that there is a distinction in the physiological or pathological status between an infant born to a non-obese and non-overweight mother and an infant born to an overweight mother as disclosed in D7.

1.6.9 At the same time, the relative fat content and absolute body fat mass have been invoked by the appellant as markers for the infant's physiological status (a pathological status does not appear to apply to healthy children).

1.6.10 The appellant's line of argument is that the group of infants born to mothers having a BMI < 25 prior to establishment of pregnancy can be taken as a new group of (treated) subjects over document D7. It was also stated on page 1, lines 25 to 27 of D7 that infants born to overweight and obese mothers had a greater risk of becoming overweight and obese later in life than infants born to lean mothers. This greater risk existed because they were thus somehow different.

1.6.11 D7 already individualises the treatment of the remaining groups of infants in relation to the mothers' BMI. The infants recited in claim 1 are first and foremost characterised by a characteristic of their mother prior to establishment of pregnancy, rather than by parameters that would render the individual infants discernible from infants forming part of said two remaining infant groups. As mentioned above, however, the group of infants born to lean mothers includes sub-collectives that have an increased risk of adiposity later in life. It is therefore not a homogeneous group of patients.

1.6.12 In view of the above, and in particular the greatly overlapping distribution curves for absolute and relative fat mass of the infants, the board sees no invoked physiological parameter that could distinguish the collective of infants born to mothers having a BMI < 25 prior to establishment of pregnancy from the remaining groups, in particular from infants born to mothers having a BMI between >= 25 and < 30 prior to establishment of pregnancy (overweight mothers). In view of the heterogeneity of the target group, it is considered impossible to establish whether a given infant fulfils the criterion "born to a non-obese and non-overweight mother", corresponding to a mother's BMI < 25 prior to establishment of pregnancy. The board thus shares the opposition division's corresponding concerns.

1.6.13 The situation underlying the patent on the relevant date differs markedly from the scenario encountered in decision T 321/15, which was cited by the appellant on appeal. In decision T 321/15 (based on European patent D7a as counterpart to D7), the deciding board held that infants born to obese mothers were distinct from infants in general. The deciding board's conclusion was, inter alia, that the group of infants born to obese mothers exhibited a pathophysiological status which differed from that of infants born to non-obese mothers (i.e. having a BMI < 30 prior to the establishment of pregnancy). No conclusion with regard to the establishment of a new group of treated subjects was drawn with regard to infants born to overweight mothers or mothers having a BMI < 25 prior to conception. In the present case, document D7 forms part of the prior art under Article 54(2) EPC. The implication of this is that, in the present case, e.g. the treated patient/subject collective "infants born to mothers having a BMI >= 25 < 30" was explicitly disclosed in the prior art. For these reasons, the reference to T 321/15 cannot support the appellant's case.

1.6.14 The case underlying decision T 694/16 is based on claimed subject-matter involving two specific biological markers and a patient population that can be determined using established criteria based on cognitive and functional tests (see points 5.5 and 5.20 of the Reasons). The patient groups considered in that decision thus do not overlap other than the claimed group having been considered to be encompassed/embedded in the larger (more generic) patient group treated in the prior art. Hence, the appellant's discussion of the overlap of the values for the clinical markers relied on in T 694/16 in the light of Figure 1c of D14a, referred to in that decision, is without merit. Moreover, the relevant marker concentration ranges appear in the claims of the relevant claim request in that case, unlike in the current case.

1.6.15 In the case underlying T 1399/04, the members of the group of subjects in question could be distinguished from all other chronic hepatitis C patients by a specific virus genotype and viral load, which can be easily determined and constitute clear biological parameters. This made it possible to differentiate members of the target group from all other chronic hepatitis C patients. This conclusion from T 1399/04 was explicitly cited by the appellant in its statement of grounds of appeal. In the present case, however, the group of subjects is determined by characteristics of the mothers prior to establishment of pregnancy. Whether or not the infants born to lean mothers constitute the largest group of subjects (as possibly the target patient group in T 1399/04) is not relevant to the board's conclusion on novelty in the current case.

1.6.16 Similarly, in the oral proceedings before the board, the appellant referred to decisions by the Boards of Appeal that dealt with infants delivered by Caesarean section as a treated patient/subject group. That group had a specific intestinal microbiome that differed from that of infants delivered by vaginal birth. The appellant argued that exactly the same scenario applied as in the current case: a physiologically different sub-group of infants was defined by the way in which they had been delivered by their mothers. There were also infants that had a microbiome that differed from that typically observed in the aforementioned group born by Caesarean section.

The board observes that the feature relied upon was not comparable as the infants could be distinguished in that case. First, the status of the infant at birth should be known and is not characterised by an event that typically occurs nine months prior to birth. By contrast, in the current case, the relevant event is the mother's BMI prior to establishment of pregnancy. Secondly, the infant's microbiome is characteristic and associated with delivery by Caesarean section. The infant's microbiome is considered a marker suitable for distinguishing between infants delivered by Caesarean section and vaginal birth, unlike absolute and relative body fat mass, which is argued to distinguish the target group in the case in hand by its average value at largely overlapping distribution curves. Therefore, this line of argument does not support the appellant's case.

1.6.17 In view of the above, novelty of the subject-matter of claim 1 has to be denied over D7 for a lack of a distinct physiological status of the group of subjects recited in claim 1.

1.7 Apart from these considerations, the aforementioned criterion ii) for establishing novelty for a known substance or composition by a second or further medical use by a new treated group of subjects is not met either. This criterion requires a functional relationship between the established new group of subjects and the effect relied upon.

1.7.1 In its grounds of appeal, the appellant underlines that it had been well established on the priority date of the patent that formula-fed infants gained more weight than breast-fed infants. Rapid weight gain during infancy was also associated with an increased risk of obesity later in life. This was stated in paragraph [0006] of the patent and in D1, page 1836, left-hand column, first paragraph of text section "Introduction". Moreover, high protein intake during infancy was associated with rapid weight gain and hence an increased risk of obesity later in life. This was demonstrated in D1 (see abstract of this document and paragraph [0009] of the patent under appeal). Surprisingly, the reduction in weight gain, observed in the clinical study underlying the patent, was much more pronounced in infants born to mothers as specified in claim 1. This was also apparent from Table 3 of the patent. In view of the above, these mothers are defined in the patent as those having a BMI < 25 prior to establishment of pregnancy.

1.7.2 However, the board notes that no data comparing the two sub-groups "mothers having a BMI < 25" and "mothers having a BMI >= 25 < 30" prior to establishment of pregnancy have been provided. Even the data presented in the patent including both groups of infants, born to obese or overweight mothers, do not include relevant statistical indications (such as the P-values and standard deviations measured in the clinical trial reported).

The appellant argued in this context that the difference in weight gain between all infants receiving either "low-protein" formula or "high-protein" formula was close to being statistically significant (see P-value of about 0.07 in Table 1 of D19, to which the appellant referred in the oral proceedings). However, no data comparing the aforementioned sub-groups are provided in the patent (or in the description of the clinical trial reported in D19; see below). This would mean comparing data sets obtained for said sub-groups of infants, born to mothers having BMIs prior to conception of < 25 and >= 25 < 30, respectively. Therefore, the appellant's argument that it was highly likely that the effect of reduced weight gain for the infants of lean mothers vs. those born to obese and overweight mothers was significant, even without providing statistical significance, is not convincing.

1.7.3 In view of the indications provided in the patent and in post-published documents D19 and D19a, the board agrees with the respondent that it is reasonable to assume that D19 describes the clinical trial underlying the data presented in the patent. The data disclosed in D19/D19b for the serum concentrations of IGF-1 (insulin-like growth factor 1), determined at an age of 6 and 12 months, are not statistically significant. In addition, these data were not broken down to consider sub-groups classified according to the BMIs of the mothers prior to pregnancy (see Table S5 in D19b). As confirmed by the appellant, the study was not designed in terms of sample size to demonstrate statistically significant differences between sub-groups established by the mothers' BMIs.

In the same way, data sets for the individual sub-groups established by the mothers' BMIs are not provided in Table 4 of the patent either. In the patent and in D19, the growth hormone IGF-1 seems to have been monitored as a predictive marker for the risk of developing obesity later in life and for infants' weight gain (see e.g. paragraph [0098] of the patent, referred to by the respondent at the oral proceedings before the board). Therefore, no relative reduction in the risk of developing the diseases recited in claim 1 has been corroborated by the appellant for the group claimed.

1.7.4 Whilst the appellant argued in favour of statistical considerations in relation to the distinction of infants by their absolute and relative fat mass, it argued that statistical significance was not necessary in relation to the aforementioned second criterion for establishing a new patient group in a second medical use.

At the same time, however, the appellant held that the purported effect of reduced weight gain in children born to lean mothers was surprising in the patent and in the statement of grounds of appeal (page 5, lines 4 to 6 and page 22, lines 9 to 11). In the latter, it is stated that, if anything, a skilled person would have expected that infants born to obese and overweight mothers would benefit most from the low-protein diet.

In view of the principle of free evaluation of evidence, the board is not bound to a specific methodology when evaluating evidence adduced. In the current case, the board observes that the purported effect of reduced weight gain in infants born to "lean" mothers vs. those born to overweight mothers on feeding a protein-reduced diet has not been supported by other convincing evidence. In view of the lack of a comparison of the relevant sub-groups, the data provided in Table 3 of the patent do not render convincing an effect that would be associated with the sub-group of infants born to lean mothers, let alone data that would be statistically relevant. There is thus no "clear trend to the naked eye in the current case", referred to in point 3.6.7 of the reasons for the decision of T 299/18 in the context of different data relating to an unrelated scenario. Hence, the appellant's reference to decision T 299/18 at the oral proceedings before the board does not support its case. Similarly, the reference to point 5.10.1 of a board's communication in a different case (T 1863/21, concerning European patent EP 2 575 508) can merely serve as a reminder that statistical significance is a concept that bears the risk of unjustly disregarding experimental results; however, as correctly pointed out by the respondent, in the current case, there is no experimental evidence comparing the target group of subjects with the infants born to mothers having a BMI >= 25 < 30.

Hence, it is also not relevant for the present decision that the opposition division had held in their preliminary opinion (D18, point 6.4), rather than in their decision, that a reduction in weight gain in the clinical trial with the protein-reduced diet was more pronounced in infants born to mothers having a BMI < 25. The opposition division referred to Table 3 of the patent in this context.

1.8 It is for these reasons that the board considers that the aforementioned criteria for establishing a new group of subjects treated have not been met. Resuming this point, it has first not been demonstrated that there is a difference between the physiological status of the known group of infants as characterised by "mother's BMI >= 25 < 30" and that of the known group recited in claim 1 ("infants born to mothers with BMI < 25"). Second, a functional relationship between the respective physiological status and the alleged therapeutic effect has not been corroborated by the appellant.

1.9 Due to the finding that the compositions and their use as specified in claim 1 are disclosed in D7 for the above-indicated reasons, the board concludes that the subject-matter of claim 1 lacks novelty over document D7 (Article 54 EPC).

2. Novelty - auxiliary requests 1 to 11

The appellant did not present any further arguments with regard to the novelty of the subject-matter of the auxiliary requests at the oral proceedings. Hence, the board's negative assessment of the novelty of these requests, outlined in its communication, is essentially repeated in the following.

The remarks made in section 1 above apply equally to the analysis of novelty of claim 1 of each of auxiliary requests 1 to 11. With regard to the additional features contained in each claim 1, the board endorses the conclusions provided in sections 5 to 9 of the Reasons of the decision under appeal. Furthermore, document D7 also discloses the feeding period from 3 to 12 months (see lines 2 to 5 on page 9; page 3, line 1; page 4, line 6), as required in claim 1 of auxiliary request 2. Hence, none of the auxiliary claim requests meets the requirement of Article 54 EPC either.