OPINION OF ADVOCATE GENERAL

JACOBS

delivered on 16 September 2004 (1)

Case C-74/03

SmithKline Beecham plc

v

Lægemiddelstyrelsen

1.     In the present case, the Østre Landsret (Eastern Regional Court) of Denmark has referred two questions concerning the Community rules governing applications to the competent authorities in the Member States for authorisation to market medicinal products. Those rules were at the material time contained in Council Directive 65/65/EEC of 26 January 1965 on the approximation of provisions laid down by law, regulation and administrative action relating to proprietary medicinal products (‘the Directive’),(2) as amended, in particular, by Council Directive 87/21/EEC of 22 December 1986.(3)

2.     Article 4.8(a)(iii) of the Directive laid down an abridged procedure for obtaining such an authorisation, whereby an applicant could rely, in support of its application, upon data submitted in respect of another previously authorised product provided, among other things, it could show that product to be essentially similar to its own. The Community case-law, considered below,(4) specified that, in order for two products to be essentially similar, they must inter alia have the same qualitative and quantitative composition in terms of active principles.

3.     The first question raised in these proceedings is whether two products which share the same active substance but in the form of different salts could be said to meet that latter criterion. The second is whether an applicant under the abridged procedure was entitled to submit additional documentation in the form of pharmacological or toxicological tests or clinical trials to a competent authority in order to show the essential similarity of its own and the previously authorised product.

4.     Since the material time, the Community rules have been amended several times and now supply clear affirmative answers to both questions.

 Legal framework

 The Directive

5.     At the material time, the relevant rules were primarily contained in Chapter II of the Directive as amended, in particular, by Directive 87/21/EEC.(5) Article 3 of the Directive provided that, in the absence of a Community-wide authorisation, a medicinal product could be marketed in a Member State only after authorisation had been obtained from the competent authority in the Member State.

6.     Article 4 defined in detail the procedure, documents and information needed in order to obtain a marketing authorisation from the competent authorities of a Member State. It laid down several possible procedural routes for obtaining a national marketing authorisation. Under the full procedure, provided for in Article 4.8, an applicant for a marketing authorisation needed ordinarily to supply the results of:

‘–      physico-chemical, biological or microbiological tests;

–       pharmacological and toxicological tests;

–       clinical trials’.

7.     Article 4.8(a) laid down several alternative procedures whereby, in certain specified circumstances, an applicant for a marketing authorisation was relieved of the obligation to provide the results of pharmacological and toxicological tests or of clinical trials as ordinarily required by Article 4.8, and was able to rely instead on data submitted in respect of another ‘reference’ product which had already been authorised. The obligation to provide full particulars of the physico-chemical nature of the product was not affected. In order to avail itself of one such procedure (hereinafter ‘the abridged procedure’), an applicant was required to demonstrate:

‘(iii) … that the medicinal product is essentially similar to a product which has been authorised within the Community, in accordance with Community provisions in force, for not less than six years and is marketed in the Member State for which the application is made; … a Member State may … extend this period to 10 years by a single Decision covering all the products marketed in its territory where it considers this necessary in the interests of public health …’.

8.     The final subparagraph of Article 4.8(a) contained the following proviso (‘the proviso’) to the procedures established by that provision:

‘However, where the medicinal product is intended for a different therapeutic use from that of the other medicinal products marketed or is to be administered by different routes or in different doses, the results of appropriate pharmacological and toxicological tests and/or of appropriate clinical trials must be provided.’

9.     The proviso thus had the effect of establishing a further procedure for obtaining marketing authorisation, often termed and hereafter referred to as the hybrid abridged procedure. Under that procedure, the applicant was required to provide only the results of such pharmacological and toxicological tests and clinical trials as were appropriate in the light of the difference in therapeutic use, route of application or dose from the other medicinal products marketed. Otherwise, the applicant relied upon the data relating to the reference product which it was required to specify under Article 4.8(a)(iii). The hybrid abridged procedure therefore represented a procedure intermediate between the abridged and the normal procedures as regards the evidential burden which it imposed on the applicant. The fresh data which an applicant was required to submit pursuant to the hybrid abridged procedure were referred to as ‘bridging data’.

10.   The preambles to the Directive and to Directive 87/21 shed light on the objectives underlying Article 4 of the Directive. The first recital of the preamble to the Directive made clear that the primary purpose underlying all the rules governing the marketing authorisation of medicinal products is the protection of public health. The second recital emphasised, however, that that objective must be obtained by means which would not hinder the development of the pharmaceutical industry or trade in medicinal products within the Community. As the third and fourth recitals indicated, disparities between national provisions would hinder trade in medicinal products and therefore needed to be removed.

11.   Directive 87/21 introduced the version of Article 4.8 which was in force at the material time. The second recital of its preamble justified the provision on the basis that ‘experience has shown that it is advisable to stipulate more precisely the cases in which the results of pharmacological and toxicological tests or clinical trials do not have to be provided with a view to obtaining authorisation for a proprietary medicinal product which is essentially similar to an authorised product, while ensuring that innovative firms are not placed at a disadvantage’. The fourth recital also drew attention to the existence of reasons of public policy for not conducting repetitive tests on humans or animals without overriding cause.

 The Notice to Applicants

12.   Guidance as to the nature of the tests and trials required in order to satisfy the various procedures laid down by Article 4 of the Directive was set out in the Annex to Council Directive 75/318/EEC of 20 May 1975 on the approximation of the laws of Member States relating to analytical, pharmaco-toxicological and clinical standards and protocols in respect of the testing of proprietary medicinal products,(6) as amended by Commission Directive 91/507/EEC of 19 July 1991.(7)

13.   The Annex required that applications for marketing authorisation be presented in a manner which took account of the guidance published by the Commission in ‘The rules concerning medicinal products in the European Community’, volume II: Notice to applicants for marketing authorisations for medicinal products for human use in the Member States of the European Community (commonly known, and hereinafter referred to, as the Notice to Applicants).

14.   Section 5 of chapter 1 of the 1998 edition of the Notice to Applicants dealt with abridged applications for marketing authorisation. Section 5.4 was entitled ‘other abridged applications’. It contained a table setting out the additional data usually required in respect of different types of abridged application. Indent (a) of the table concerned applications where the product for which authorisation is sought has a ‘different salt/ester complex/derivative (with the same therapeutic moiety)’, the therapeutic moiety being that portion of the product which is primarily responsible for its therapeutic effect upon the consumer. It identified the following additional data as necessary:

‘Evidence that there is no change in the pharmacokinetics of the moiety, pharmacodynamics and/or in toxicity which could change the safety/efficacy profile (otherwise, to be considered as a new active substance)’.

15.   In versions of the Notice to Applicants which have been issued since May 2001 (and therefore after the applications for marketing authorisation at issue in the present proceedings), the following explanation is included as to the meaning of a new active substance:

‘A definition on what has to be understood by new active substance is given in Appendix III of this chapter.

Active substance(s) in the form of different salts, esters, derivatives, etc. but with the same active therapeutic moiety should (normally) not be considered as a new active substance, unless they differ significantly from each other in properties regarding safety or efficacy. Evidence on same properties regarding safety and efficacy of a different salt, ester, other derivative of an active substance should be provided by the applicant when he claims that the medicinal product for which the application is submitted is essentially similar to an authorised medicinal product containing another salt, ester or other derivative of the same active substance. (See Appendix IV of this chapter).

The decision whether a different form of the active substance is to be classified as a new active substance should be taken by the competent authorities of the Member States/EMEA on a case-by-case basis.’

16.   Appendix IV to chapter 1 of volume 2A reproduces indent (a) of the table previously contained in section 5.4 of the 1998 version of the Notice to Applicants, set out above.(8)

 The Generics judgment

17.   In the Generics case,(9) the Court considered the meaning of essential similarity for the purposes of Article 4.8(a). It emphasised that the requirement of essential similarity must be interpreted in the light of the Directive’s primary purpose of safeguarding public health, so as to avoid any relaxation of the requirement of safety and efficacy imposed upon all medicinal products.(10)

18.   The Court proceeded to hold that one medicinal product is essentially similar to another ‘where it satisfies the criteria of having the same qualitative and quantitative composition in terms of active principles, of having the same pharmaceutical form and of being bioequivalent, unless it is apparent in the light of scientific knowledge that it differs significantly from the original product as regards safety or efficacy’. The Court also stated that the competent authority of a Member State may not disregard the three criteria identified when it is required to determine whether two products are essentially similar.(11)

19.   The Court considered that the criteria of qualitative and quantitative composition in terms of active principles, pharmaceutical form and bioequivalence would help to ensure that a new product would have the same safety and efficacy as its reference product. However, the Court still considered it necessary to add the final condition that the two products should not otherwise significantly differ in terms of their safety and efficacy. It illustrated the need for the final condition by noting that ‘a medicinal product which satisfies the three criteria referred to … may nevertheless raise questions of safety related to its excipients’.(12)

 Subsequent amendments

20.   Since the material time, the Directive has been replaced by Directive 2001/83 (‘the new Directive’).(13) For an initial period, the abridged procedure, now in Article 10(1)(a)(iii) of the new Directive, remained in substance the same as the version applicable at the material time. During that time, Annex I to the new Directive was amended by Commission Directive 2003/63/EC of 25 June 2003.(14) Part II of the Annex dealt with ‘specific marketing authorisation dossiers and requirements’. At point 2 thereof, it was stated that in order to demonstrate essential similarity, applications under the abridged procedure should contain, ‘if applicable, additional data in order to demonstrate evidence on the equivalence of safety and efficacy properties of different salts, esters or derivatives of an authorised active substance’. Similarly, it was stated at point 3 that:

‘Where the active substance of an essentially similar medicinal product contains the same therapeutic moiety as the original authorised product associated with a different salt/ester complex/derivative evidence that there is no change in the pharmaco-kinetics of the moiety, pharmaco-dynamic and/or in toxicity which could change the safety/efficacy profile shall be demonstrated. Should this not be the case, this association shall be considered as a new active substance.’

21.   The abridged procedure was then significantly recast by Directive 2004/27/EC of 31 March 2004. (15) Under Article 10(1) of the new Directive as amended, the concept of essential similarity has been replaced by a (basically analogous) requirement upon applicants to show that their product is a generic of a reference medicinal product which has been authorised for not less than eight years. Article 10(2)(b) defines a generic medicinal product as:

‘… a medicinal product which has the same qualitative and quantitative composition in active substances and the same pharmaceutical form as the reference medicinal product, and whose bioequivalence with the reference medicinal product has been demonstrated by appropriate bioavailability studies. The different salts, esters, ethers, isomers, mixtures of isomers, complexes or derivatives of an active substance shall be considered to be the same active substance, unless they differ significantly in properties with regard to safety and/or efficacy. In such cases additional information providing proof of the safety and/or efficacy of the various salts, esters or derivatives of an authorised active substance must be supplied by the applicant.’

 Facts

22.   In the present case, SmithKline Beecham plc (hereinafter ‘SmithKline Beecham’), a pharmaceuticals undertaking, challenges the validity of marketing authorisations granted by the Lægemiddelstyrelsen, which is the competent authority in Denmark, to two other pharmaceuticals undertakings, Synthon BV and Genthon BV (hereinafter ‘Synthon and Genthon’), both of which belong to the Synthon group.

23.   SmithKline Beecham is the current holder of a marketing authorisation in respect of the medicinal product, Seroxat. The active substance in Seroxat takes the form of paroxetine hydrochloride hemihydrate in dosages of 20 mg and 30 mg. The first marketing authorisation for Seroxat dates from 1993.

24.   In July 1999, Synthon and Genthon submitted largely identical applications for authorisation to market Paroxetine ‘Synthon’ and Paroxetine ‘Genthon’ respectively (hereinafter, the ‘Synthon product’). Their applications were made pursuant to the abridged procedure, citing Seroxat as the reference product. Like Seroxat, the Synthon product contains paroxetine, but in the form of a different salt, paroxetine mesylate.

25.   In addition to the documentation which they were required to submit under the abridged procedure, Synthon and Genthon also supplied results from selected pharmacological and toxicological tests on animals of the type referred to in the Annex to Directive 75/318 as amended by Directive 91/507. The Lægemiddelstyrelsen subsequently requested further information.

26.   Synthon and Genthon did not submit results of clinical trials on patients but requested that an exception be allowed on the grounds that the effects of their product on humans were documented indirectly by way of bioequivalence to the reference product, Seroxat, tested on a number of healthy human volunteers.

27.   In October 2000, the Lægemiddelstyrelsen granted marketing authorisations to Synthon and Genthon in respect of the Synthon product.

 National proceedings and questions referred

28.   SmithKline Beecham has brought proceedings before the Østre Landsret to contest the legality of the Lægemiddelstyrelsen’s decision to authorise the Synthon product.

29.   It argues that Seroxat and the Synthon product are not essentially similar to one another because they contain different, albeit related, active substances. The active substance in Seroxat is paroxetine hydrochloride hemihydrate whereas that in the Synthon product is paroxetine mesylate, another paroxetine salt. Under the test of essential similarity, it should be immediately apparent whether two products are essentially similar or not. The fact that further data were necessary in order to demonstrate essential similarity is sufficient to confirm that the active substances in Seroxat and the Synthon product are different. The submission of further particulars in the form of pharmacological and toxicological tests or clinical trials in the context of an abridged procedure is permitted only pursuant to the proviso, namely where the new product is intended for a different therapeutic indication, or is to be administered by different routes or in different doses.

30.   The Østre Landsret has decided to stay the proceedings before it and to refer the following questions to the Court of Justice:

‘(1)      Is it compatible with point 8(a)(iii) of Article 4 of the First Medicinal Products Directive (Directive 65/65/EEC, as amended) for a product to be authorised under the abridged application procedure when a salt of the active substance in the product is changed from the one used in the reference product?

(2)      Can the abridged application procedure be used when an applicant, on its own initiative or at the request of national health authorities, submits additional documentation in the form of certain pharmacological or toxicological tests or clinical trials with a view to demonstrating that the product is “essentially similar to” the reference product?’

31.   Written observations have been received from SmithKline Beecham, Synthon and Genthon, the Danish, Netherlands, Portuguese and United Kingdom Governments, and from the Commission, all of whom, with the exception of the Netherlands and Portuguese Governments, made oral submissions at the hearing.

 Assessment

32.   It is clear from the order for reference and from the submissions before the Court that the two questions referred are closely connected. One of the arguments against finding that two products can be essentially similar despite a change in the salt of the active substance (the issue raised in the first question) is that the applicant would generally then need to submit additional data in order to show that they did not differ in terms of their safety and efficacy, which is argued to be impermissible (the issue raised in the second question). I therefore propose to consider the two questions together.

33.   Synthon and Genthon, Denmark, the Netherlands and the Commission support an affirmative response to the questions referred, as, in effect, does Portugal. SmithKline Beecham and the United Kingdom submit that both questions should be answered in the negative.

34.   Plainly, the text of the Directive does not in itself supply a solution to the questions referred. The meaning of essential similarity is not defined. Nor is it entirely clear whether any limits attach to the additional data which may be submitted in the context of an abridged application.

35.   The Court’s case-law is similarly open to interpretation. The judgment in Generics does not specify in any detail what is meant by ‘the same qualitative and quantitative composition in terms of active principles’. It appears from the submissions before the Court that the term ‘active principle’ is used to refer both to the whole molecule constituting the active ingredient and, where that ingredient is in the form of a salt, to the portion of the molecule (often referred to as the ‘therapeutic moiety’) which has a therapeutic effect upon the consumer of the medicine.

36.   As regards the data which an applicant may submit, the case-law has hitherto been more concerned with the circumstances in which an applicant can avoid supplying data by relying on those previously provided in support of another product.

37.   As a consequence, the submissions before the Court have for the most part focussed more generally upon the scheme of the Directive; the objectives which underlie it; and the relevance of other provisions of Community law and of the Notice to Applicants.

 The scheme of the Directive

38.   The United Kingdom and SmithKline Beecham emphasise the importance to the scheme of the Directive of the distinction between the abridged procedure provided for in Article 4.8(a)(iii) and the hybrid abridged procedure established by the proviso. In their submission, that distinction would be undermined if the definition of essential similarity, adopted by the Court in Generics, were to be relaxed and the routine submission of additional data were allowed in a wider set of circumstances than those explicitly or implicitly included within the proviso.

39.   They submit that the Court’s definition of essential similarity should be interpreted in such a way that, when the criteria identified are met, it can safely be assumed that the two products being compared will have the same safety and efficacy profile. The final condition, namely that the two products should not differ substantially in terms of their safety and efficacy, should be employed only as an additional safeguard, or ‘fall-back’ question, against the risk that a change in the excipients used might render the new product less safe or efficacious.

40.    However, if two salts sharing the same therapeutic moiety are deemed to constitute the same active substance, that criterion will no longer be a good measure of equivalence in terms of safety and efficacy. The substitution of one salt for another could affect therapeutic efficacy by improving or reducing the absorption of a drug and its bioavailability, or could affect its toxic potential or stability, resulting in adverse effects.

41.   As a result, an applicant would ordinarily have to submit additional data in order to demonstrate that, despite the change of salt form, no significant difference arose as to the safety and efficacy of the two products under consideration. The final condition of the Court’s test in Generics would need to be deployed in most, if not all, such situations. It would cease to be discrete from the active substance criterion but would in effect replace it. Moreover, an applicant would be enabled to submit bridging data in a wider range of circumstances than are identified (explicitly or implicitly) by the proviso.

42.   I am not convinced by those submissions.

43.   Given the wording of the second question, which clearly arose out of the present line of argument, it is necessary to state at the outset that there is to my mind no general rule against the submission of additional data in the context of the abridged procedure. An affirmative answer should therefore, in my view, be given to that question.

44.   That follows from the text of the relevant provisions. The abridged procedure is stated to be available if the applicant can demonstrate, among other things, essential similarity. The clear implication is that the applicant may submit all of the evidence required for such a demonstration.

45.   Similarly, there is nothing in the proviso to suggest that additional data may only be supplied pursuant to that provision. In principle, data submitted in the context of the proviso and those provided under Article 4.8(a)(iii) serve different ends. The former are designed to compensate for a lack of essential similarity, whereas the latter go to prove the existence of essential similarity.

46.   It would also appear from the test adopted by the Court in Generics that additional data may sometimes be needed in order to show essential similarity. The Commission submits, for example, that an applicant under the abridged procedure may need to supply documentation in order to show the bioequivalence of its product with the reference product. Additional data may also be needed to show that two products which otherwise meet the criteria specified by the Court in Generics do not differ in their safety and efficacy as a result, for example, of differences in the excipients used, or (in an example given by Synthon and Genthon) of a change in the route of synthesis used in manufacturing the active ingredient.

47.   As a general proposition, I think it would be highly unfortunate and quite contrary to the scheme of the Directive to place limits upon the data which may be supplied by an applicant under the abridged procedure, either at the request of a competent authority or on its own initiative. Subject to the specific policy of data exclusivity during the stipulated period, the appropriate policy, apparent in many aspects of the relevant Community rules, is to maximise the information available to a competent authority when deciding whether to grant an authorisation.

48.   Thus, the relevant issue appears to me to be not whether additional data may be submitted in the context of the abridged procedure, but rather whether, if two salts having the same therapeutic moiety were deemed to constitute the same active substance, additional data would too often be required under the abridged procedure, thereby blurring the distinction between that procedure and the hybrid abridged procedure.

49.   I agree that there is a risk of extending too far the criteria of essential similarity, thereby placing too much reliance on the final condition of the Court’s test in Generics, and in the process eroding the distinction between the two procedures, contrary to the scheme of the Directive.

50.   However, I am not convinced that the interpretation of essential similarity supported by Synthon and Genthon, Denmark, the Netherlands, Portugal and the Commission would unduly widen the circumstances in which two products may be essentially similar to one another.

51.   On the contrary, it would appear to me an excessively narrow interpretation of essential similarity to require products to have active ingredients with exactly the same molecular structure before they can be classed as essentially similar.

52.   The text of the Directive does not specify that two products be identical in order for the abridged procedure to operate, only that they be essentially similar. Some types of difference between the two products are therefore clearly permissible.

53.   Nor does the need for an exact molecular match between active ingredients follow from the criterion, identified by the Court in Generics, that essentially similar products have the same quantitative and qualitative composition in terms of active principles. As stated above, it appears that the term active principle is sometimes used to refer to the therapeutically active moiety of an active substance as well as to the substance itself.

54.   There is a broad consensus among the parties that it is more realistic to focus on therapeutic action than on the precise molecular structure of active ingredients when considering whether two products are capable of essential similarity.

55.   Thus, the Commission submits that the therapeutic action of a medicinal product in the form of a salt typically depends upon the active moiety of a salt and not the appended portions of the molecule.

56.   Similarly, the Netherlands distinguishes, when the active ingredient of a medicinal product is in salt form, between the positive part of that salt, which it describes as the active principle, and the negative part, which it terms the inert element. When dissolved, the two parts of the salt are separated. The positive part gives the medicine its effect.

57.   Synthon and Genthon submit, as regards the products at issue in the main proceedings (namely, Seroxat and the Synthon product), that both depend upon the positive part of the salt which is their active ingredient to achieve their therapeutic effect on a patient. The negative part of the salt, which differs as between the two products, is merely an inert working vehicle to enable the products to be manufactured in tablet form. In both cases, the positive part is absorbed into the bloodstream and distributed and metabolised throughout the body. Conversely, the inert part of each product passes through the gastrointestinal tract without being absorbed and without having any effect on the body.

58.   At the hearing, the Danish Government also placed the emphasis upon what happens to the active ingredient after it has been administered to a patient. In the case of a salt, typically the two parts of the salt will separate. The therapeutically active part will be absorbed, whilst the other part is eliminated by the body.

59.   Those submissions suggest that an analogy may frequently be drawn between the non-therapeutically active portion of the salt, where an active ingredient is in the form of a salt, and any other inert constituent of a medicinal product. In so far as such an analysis is correct, it would appear unreasonable to rule out essential similarity in all cases where that portion is changed but the therapeutically active moiety remains the same as between two products. Admittedly, as in the case of excipients, there is always a risk that an alteration in the appended portion of the salt will result in a change to the safety and efficacy of the product. However, the final condition of the Generics test serves to guard against just such a risk.

60.   I therefore consider that it would not distort the scheme of the Directive if two products having active ingredients which share the same therapeutic moiety were understood to share the same qualitative and quantitative composition in terms of active principles.

61.   In order to confirm whether such an interpretation is indeed correct, however, it is necessary also to consider whether it accords with the objectives underlying the marketing authorisation regime and with the various other provisions of Community law which have been raised in argument before the Court.

 The objectives underlying the marketing authorisation regime

62.   As emerges from the recitals of the Directive, and as the Court has held, four objectives are of relevance when interpreting the Community rules governing the issuing of marketing authorisations in the Member States: the protection of public health; the avoidance of unnecessary testing on humans and animals; the protection of the interests of innovative pharmaceutical companies; and the promotion of the internal market in medicinal products, not least by ensuring a harmonised approach on the part of the competent authorities in the Member States.

 The protection of public health

63.   As regards the primary objective of protecting public health, which must underlie all legislation governing the authorisation of medicinal products, the United Kingdom and SmithKline Beecham both point to circumstances in which differing salt forms of the same active ingredient may vary in terms of their safety and efficacy.

64.   It does not appear to me, however, that risk to public health would eventuate if two salt forms having the same therapeutic moiety were found to be capable of essential similarity. The consequences of a change in salt form, identified by the United Kingdom and SmithKline Beecham, do not prove otherwise.

65.   In so far as those consequences relate to bioavailability, they will in any event preclude a finding of essential similarity, given that bioequivalence must also be shown. Where the consequences concern stability and toxicity, they will result in a difference in safety and efficacy between the products, and will be caught by the final condition of the test laid down by the Court in Generics.

 Avoidance of unnecessary testing

66.   As the Commission, Denmark, Synthon and Genthon submit, the objective of avoiding the unnecessary testing of humans and animals clearly militates in favour of an interpretation of essential similarity which could in principle encompass differing salt forms sharing the same therapeutic moiety. Rather than needing to compile an entirely new dossier of data, with the various tests which that would entail, the applicant is able to rely in large part upon the data submitted in respect of the reference product.

 Ensuring that innovative pharmaceutical companies are not placed at a disadvantage

67.   The United Kingdom submits that the test of essential similarity should be strict in order to maintain a fair balance between the interests of innovator companies and generic companies which the Directive and the judgment of the Court in Generics were intended to achieve.

68.   I do not agree.

69.   The protection of the interests of innovative pharmaceutical companies is to my mind primarily reflected in the six or ten year period of data exclusivity which is provided for under Article 4.8(a)(iii). The requirement of essential similarity serves first and foremost to safeguard public health.

70.   Moreover, the Commission points to the possibility that, if different salt forms having the same therapeutic moiety could never be regarded as essentially similar, pharmaceuticals companies might be able artificially to extend the period of data exclusivity which they enjoy by withdrawing a product from the market shortly before the expiry of its period of data exclusivity and launching another version of the product having the same active principle but in the form of a different salt.

71.   To the extent that such a risk does indeed exist, it would appear to be reduced by the interpretation of essential similarity favoured by the Commission, Denmark, the Netherlands, Synthon and Genthon.

      

 The internal market in pharmaceutical products and the mutual recognition procedure

72.   According to SmithKline Beecham, the notion of essential similarity must be given a meaning which can be put into operation by the competent authorities of the Member States without hesitation, in a simple and straightforward manner, thereby ensuring a harmonised approach across the Community. A common approach is particularly important given the mutual recognition procedure, whereby an applicant who obtains a marketing authorisation in one Member State may on that basis obtain authorisations in other Member States without needing in each case to make a full application. Because of their lack of equivalence in terms of safety and efficacy, related but different salts require a more complex assessment, which must be conducted on a case by case basis. If products containing such salts were none the less considered capable of essential similarity, the risk arises of a divergence of approach between competent authorities and with it the possibility of forum shopping by pharmaceuticals companies.

73.    I agree that the test of essential similarity should be as clear and objective as possible. The second recital of the preamble to Directive 87/21 indicates the legislature’s intention to stipulate more precisely the cases in which the full complement of data need not be provided by the applicant for a marketing authorisation. The identification of easily applicable standards is obviously conducive to the smooth operation of the mutual recognition procedure and therefore to the free circulation of pharmaceutical products.

74.   That said, it seems to me that some margin of discretion for the competent authorities of the Member States when considering applications pursuant to the abridged procedure is unavoidable, as is evident from the final condition of the test laid down by the Court in Generics. Competent authorities will need to take a view whether, for example, a change in the excipients as between a new product and the reference product will lead to any significant differences as regards safety and efficacy.

75.   The negative consequences for the mutual recognition procedure should not be overstated. Under any of the procedures for obtaining a marketing authorisation, some risk arises that competent authorities in the Member States will disagree as to the appropriateness of a given authorisation. Precisely because of the irreducible element of discretion involved in the authorisation regime, the Community rules lay down a system for resolving such disagreements. (16) That system appears to me fully sufficient to meet concerns of divergence and forum shopping.

76.   My overall conclusion as regards the objectives of the marketing authorisation regime is therefore that the goal of avoiding unnecessary testing on humans and animals favours the findings that two products may be essentially similar where they have the same therapeutic moiety but take the form of different salts. None of the other objectives requires a different conclusion.

 Other provisions of Community law

77.   Various aspects of the Community rules applicable in the pharmaceuticals sector are argued by SmithKline Beecham and the United Kingdom to provide indirect support for the proposition that different salt forms of an active substance can never be essentially similar.

78.   First, reference is made to Part 1 of the Annex to Directive 75/318. It is argued that the manner in which the qualitative composition of a medicinal product is described there implies that the active ingredient of a medicinal product should be understood in the case of salts to include both the active moiety and the appended portion of the molecule, and should be identified as such.

79.   It is true that Advocate General Ruiz-Jarabo Colomer did state, when proposing in Generics the criteria for essential similarity, that what was meant by the composition of a medicinal product was clearly described in the Annex. (17) However, his reason for referring to the Annex was to demonstrate that neither the excipients nor the external covering of a medicinal product was relevant to its composition.

80.   For my own part, I do not think that the Annex can be conclusive for present purposes. It served at the material time to set out the particulars and documents which must accompany applications for marketing authorisation. In that context, it makes sense that the qualitative composition of the active ingredient should be exhaustively described. Different considerations apply when interpreting the criteria of essential similarity specified by the Court in Generics.

81.   The United Kingdom and SmithKline Beecham also refer to Commission Regulation (EC) No 541/95 of 10 March 1995 concerning the examination of variations to the terms of a marketing authorisation granted by a competent authority of a Member State. (18) At the material time, that Regulation laid down the procedure for the examination of applications for variations to the terms of a marketing authorisation of medicinal products. It distinguished between minor variations, which could be effected by way of amendment to an existing authorisation, and major variations, which required an application for a new marketing authorisation to be made. The former types of variation were identified in Annex I to the Regulation and the latter in Annex II. Among the types of variation identified in Annex II were ‘changes to the active substance(s)’, which were specified to include ‘replacement of the active substance(s) by a different salt/ester complex/derivative (with the same therapeutic moiety)’. It is argued that it would be odd if a change were sufficiently important to require a new marketing authorisation but had no effect upon the conclusion that a generic product had the same qualitative and quantitative composition as a reference product.

82.   As the United Kingdom notes, the Court specifically ruled in Generics that Annex II to Regulation No 541/95 is of no relevance whatsoever to the application of Article 4.8(a)(iii) of the Directive. (19)

83.   In any event, I have no difficulty in accepting that a variation between two products might be deemed sufficiently important to necessitate an application for a new marketing authorisation under Regulation No 541/95 whilst at the same time falling within the scope of one or other of the procedures which were at the material time laid down by Article 4.8(a). It is noteworthy that many of the other types of change identified in Annex II would fall within the hybrid abridged procedure provided for in the proviso to Article 4.8(a).

84.   Finally, SmithKline Beecham seeks to derive support for its reading of essential similarity from the Community rules relating to orphan medicinal products which serve to diagnose, prevent or treat rare illnesses and which it would therefore be uneconomic to develop under normal market conditions. Article 8 of Regulation (EC) No 141/2000 of the European Parliament and of the Council of 16 December 1999 on orphan medicinal products (20) confers upon such products a ten-year period of marketing exclusivity, during which no authorisation is to be granted in respect of any similar medicinal product for the same therapeutic indication.

85.   Article 3 of Commission Regulation (EC) No 847/2000 of 27 April 2000 (21) defines a similar medicinal product to mean ‘a medicinal product containing a similar active substance or substances as contained in a currently authorised orphan medicinal product and which is intended for the same therapeutic indication’. A similar active substance is in turn defined as ‘an identical active substance, or an active substance with the same principal molecular structural features (but not necessarily all of the same molecular structural features) and which acts via the same mechanism’. It is defined to include ‘isomers, mixture of isomers, complexes, esters, salts and non-covalent derivatives of the original active substance, or an active substance that differs from the original active substance only with respect to minor changes in the molecular structure, such as a structural analogue’.

86.   The fact that two products are considered to be similar rather than identical when they take the form of different salts is said to illustrate that such products cannot have the same qualitative and quantitative composition in terms of active principles for the purposes of the test of essential similarity identified in Generics.

87.   In my view, the definitions contained in Regulation No 847/2000 are very specific to their context and are not relevant when interpreting Article 4.8(a)(iii) of the Directive. No one, I think, would claim that two different salts which have the same therapeutic moiety are identical. The question for present purposes is whether they can be said to have the same qualitative and quantitative composition in terms of active principles.

88.   In any event, even assuming that Article 3 of that regulation were relevant, a definition of a similar active substance is expressed to include an identical active substance, and can therefore hardly be used as a basis for distinguishing between similar and identical substances.

89.   Hence, I do not consider that any of the various legislative provisions identified by the United Kingdom and SmithKline Beecham run counter to my conclusion that two products which share the same therapeutic moiety but in different salt forms may be essentially similar for the purposes of Article 4.8(a)(iii) of the Directive.

 The Notice to Applicants

90.   The Commission, Denmark, the Netherlands, Synthon and Genthon all refer to the Notice to Applicants in support of the proposition that two different salts which contain the same therapeutic moiety may be essentially similar within the meaning of Article 4.8(a)(iii) of the Directive.

91.   In response, the United Kingdom and SmithKline Beecham emphasise first that the document is not legally binding. SmithKline Beecham also notes that the 2001 version of the Notice to Applicants cannot be relevant in the present proceedings, which were commenced prior to its publication. In the United Kingdom’s submission, even the 1998 version of the Notice which was in force at the material time does not provide unambiguous support for the proposition that two salts with the same therapeutic moiety might in principle be essentially similar. That version states that such salts will not be essentially similar unless evidence is submitted to show that ‘there is no change in the pharmacokinetics of the moiety, pharmacodynamics and/or in toxicity which could change the safety/efficacy profile’. According to the United Kingdom, there is always an irreducible possibility that a change of salt form could affect safety and efficacy. The 1998 version of the Notice is therefore consistent with the view that two products having different salts may never be regarded as essentially similar.

92.   For the reasons set out in my Opinion in Approved Prescription Services, (22) I consider that while that Notice is not legally binding, it must be accorded some weight in the interpretation of the Directive. That is especially the case where, as here, complex technical questions are at issue.

93.   In my view, despite the United Kingdom’s suggestion to the contrary, the Notice to Applicants can only be read as supporting the view that two different salt forms having the same therapeutic moiety may qualify as essentially similar. The Notice would surely not identify a possibility which it considered invariably to be foreclosed.

94.   That said, I am not totally convinced by the analysis adopted in the Notice to Applicants to arrive at that result. The Notice suggests that the issue whether two such forms constitute the same active ingredient will depend upon their safety and efficacy. It seems to me that it would have been more consistent with the test of essential similarity laid down by the Court in Generics to describe the products in question as having the same qualitative and quantitative composition in terms of active principles whilst noting that their essential similarity would none the less ultimately depend upon their being shown not to differ in terms of their safety and efficacy, in accordance with the final condition of that test.

95.   Similarly, the Notice to Applicants classifies an application made on the basis of the essential similarity of two different salt forms under the heading of ‘other abridged applications’, together with applications made under the proviso. To my mind, however, such an application should have proceeded under the standard abridged procedure. Any additional data submitted pursuant to the latter procedure would serve to demonstrate the essential similarity of the two products, not to compensate for the lack of such similarity.

96.   Whilst the Notice to Applicants is thus consistent with the result arrived at here, the manner in which it presents that conclusion may be responsible for some of the uncertainty and confusion which have arisen in the context of the present proceedings.

97.   In the light of the scheme and objectives of the Directive, the other provisions of Community law considered above, as well as the Notice to Applicants, I am thus of the view that two products containing different salt forms of the same active principle were indeed capable of being essentially similar under the abridged procedure as it operated prior to its recent amendments. It was also open to an applicant under that procedure, either at the request of the competent authority in a Member State or on its own initiative, to submit additional data in order to demonstrate essential similarity.

98.   I note that the conclusion which I have reached here matches the interpretation of essential similarity which was explicitly endorsed in the Annex to Directive 2001/83 (23) following its amendment by Directive 2003/63. (24) It seems at least arguable, from the manner in which the relevant provisions of the replacement Annex were framed, that they were intended to record the Commission’s understanding of the existing legal position rather than to amend the status quo.

99.   I also note that the Community legislature has since chosen explicitly to revalidate such an approach in the new version of the abridged procedure, introduced by Directive 2004/27. (25) Clearly, earlier versions of the Directive fall to be interpreted independently of such subsequent developments. None the less, the legislature’s choice does at least suggest that the conclusion arrived at here was considered to be workable and appropriate in the light of the objectives of the Community rules relating to marketing authorisation.

 Conclusion

100. I am therefore of the opinion that the Court should respond as follows to the questions referred by the Østre Landsret:

(1)      Two products which share the same therapeutic moiety but take the form of different salts have the same qualitative and quantitative composition in terms of active principles and may therefore be essentially similar for the purposes of Article 4.8(a)(iii) of Council Directive 65/65/EEC of 26 January 1965 on the approximation of provisions laid down by law, regulation or administrative action relating to proprietary medicinal products, as amended, provided that they do not differ significantly as regards safety or efficacy.

(2)      In support of an application made pursuant to that provision, an applicant may, on its own initiative or at the request of the competent authority in a MemberState, submit additional documentation in the form of pharmacological or toxicological tests or clinical trials in order to demonstrate that its product is essentially similar to the reference product.

---

1 – Original language: English.

---

2 – OJ, English Special Edition 1965-1966, p. 24.

---

3 – OJ 1987 L 15, p. 36.

---

4  – At paragraphs 17 to 19.

---

5 – The Community legislative framework for medicinal products was with effect from 18 December 2001 codified and consolidated in Directive 2001/83/EC of the European Parliament and of the Council of 6 November 2001 on the Community code relating to medicinal products for human use, OJ 2001 L 311, p. 67. Subsequent amendments to the relevant provisions of that directive are set out below.

---

6 – OJ OJ 1991 L 270, p. 32.

---

7 – OJ 1991 L 270, p. 32.

---

8 – At paragraph 14

---

9 – Case C-368/96 Generics (UK) and Others [1998] ECR I-7967.

---

10  – Paragraph 22.

---

11  – At paragraphs 36 and 37 and in the operative part of the judgment.

---

12  – Paragraph 32.

---

13  – Cited in note 5.

---

14  – OJ 2003 L 159, p. 46.

---

15  – OJ 2004 L 136, p. 34.

---

16  – The relevant provisions are now contained in chapter 4 of Directive 2001/83.  At the material time, they were to be found in the Directive and in Directive 75/319/EEC of 20 May 1975 on the approximation of provisions laid down by law, regulation or administrative action relating to proprietary medicinal products, OJ 1975 L 147, p. 13, as those directives were amended, in particular, by Directive 93/39/EEC of 14 June 1993, OJ 1993 L 214, p. 22.

---

17  – At paragraph 37 of the Opinion.

---

18  – OJ 1995 L 55, p.7.

---

19  – At paragraph 58 of the judgment.

---

20  – OJ 2000 L 18, p. 1.

---

21  – OJ 2000 L 103, p. 5.

---

22  – Case C-36/03, delivered on 8 July 2004, paragraphs 70 to 73.

---

23  – Cited in note 5.

---

24  – Cited in note 14.

---

25  – Cited in note 15.