First inclusion of a racemate in the ARTG also constitutes the first inclusion of a “pure” enantiome

First inclusion of a racemate in the ARTG also constitutes the first inclusion of a “pure” enantiome

Alphapharm Pty Ltd v Lundbeck A/S [2008] FCA 559

The Federal Court has recently handed down a decision which will be of particular interest to the pharmaceutical industry. At issue was whether a single enantiomer can be patented in light of the prior disclosed racemate. In concurrently heard proceedings Lindgren J also considered the question of whether a patent directed to a purifed enantiomer is eligible for a patent term extension, and also, whether data submitted to the TGA concerning the purified enantiomer is “protected information”, where the racemate, containing the enantiomer, had previously been included in the ARTG.

Facts of the case

Australian Patent No 623144 (“the patent“) to H Lundbeck A/S (“Lundbeck“) was filed on 13 June 1989 and relates specifically to the (+)-enantiomer of citalopram, the (+)enantiomer also referred to as escitalopram. Citalopram is a selective serotonin reuptake inhibitor (SSRI) and is used in the treatment of depression. It contains a single chiral centre, which affords two possible enantiomeric forms, designated as the (+) and (-) forms. These refer to the individual enantiomer’s ability to rotate plane polarised light to the right and left respectively. A mixture comprising an equal measure of both enantiomers is referred to as the racemate, or a racemic mixture, and results in no net rotation of plane polarised light. Racemic citalopram was disclosed in Lundbeck’s earlier Australian Patent No 5509445 (“the ‘445 patent”). The (+)enantiomer was described in the later patent as being therapeutically more active than the racemate and the (-)-enantiomer.

The patent contained six claims, being directed to the specific (+)-enantiomer per se, addition salts thereof, pharmaceutical compositions comprising the (+)-enantiomer in unit dosage form, as well as to methods for the preparation of the compound. The 20 year term is set to expire on 13 June 2009. Escitalopram oxalate was included on the ARTG on 16 September 2003 and on 22 December 2003 Lundbeck applied for an extension of term of the patent. An extension to 13 June 2014 was granted and particulars were entered into the Register. In 2006, a Delegate of the Commissioner held that the first regulatory approval date was in fact earlier than that alleged by Lundbeck and directed, under reg.10.7(7), that the Register be amended to reflect an extension some 18 months shorter, to 9 December 2012.



Alphapharm’s attack on the novelty of the claims centred on the construction of the patent. In particular, Alphapharm contended that reference in the claims to the (+)enantiomer, without some qualification of being “individual” or “pure”, did not import a level of optical purity over the racemate, and merely implied reference to that enantiomer as present in the racemate. However, taking into consideration what was the convention and common general knowledge at the priority date (1988) Lindgren J favoured an interpretation which required at least 95% pure (+)-citalopram. His Honour readily concluded that unqualified reference to the (+)-enantiomer of citalopram in claim 1 is specific reference to the enantiomer that rotates plane polarised light to the right, and refers to something different from that enantiomer as an indistinguishable part of the unresolved racemate. His Honour noted that this conclusion was further supported by numerous references to “individual”, “pure”, “isolated” and “separated” throughout the body of the specification which as a whole teach that claim 1 refers to the independently existing enantiomer.

Having established how the claims are to be construed, the question of novelty essentially turned on whether the racemate disclosed in the earlier ‘445 patent anticipated claims to the isolated (+)-enantiomer. His Honour acknowledged that the skilled addressee would have understood that the racemate consisted of the (+) and (-) isomers in equal parts, and would have been able to identify the (R)- and (S)- designated forms,, although in the absence of further experimentation, would not have known which correlated with the (+) and (-) forms. His Honour found that the ‘445 patent made no reference to enantiomers, let alone expressly or by implication disclosed the individual enantiomers. Given that the construction placed on the claims required the independent existence of the (+)enantiomer, the claims were held to be novel over the ‘445 patent. An earlier publication suggesting that (-) citalopram to be the more potent of the two isomers, whilst recognized as disclosing an individual enatiomer, was considered to actually teach away from the invention and therefore also was not considered to deprive the claims of novelty.

The question of obviousness was acknowledged to be conceptually different for the product claims and the method claim. For the product claim, the question considered was whether the goal of the (+)enantiomer is an obvious one to pursue. For the method claims, the starting point assumes the goal of seeking to obtain the (+)-enantiomer and the question is whether the means for achieving that goal are obvious. However, it was noted that the question of obviousness or non-obviousness of the method claim was potentially relevant to the enquiry concerning the product claims. His Honour acknowledged that it was part of the common general knowledge that a racemate consisted of the (+) and (-) enantiomers and that resolution of racemic mixtures was known. Nonetheless, a finding of obviousness required something more than merely conceiving the possibility of resolving the separate isomers.

The method claimed included two possible routes, (a) and (b). Alphapharm focussed its arguments on route (b), which required the use of a chiral precursor diol which was resolved and subsequently cyclised to the final product. The evidence showed that previous extensive attempts by Lundbeck at resolving the individual citalopram enantiomers, through commonly used methods such as chromatography or crystallization of diastereomeric salts, had proved unsuccessful. Furthermore, a route which required resolution of the precursor diol was initially dismissed by Lundbeck for a number of reasons, including the expectation that the subsequent ring closure required to form citalopram would be expected to result in racemisation. His Honour accepted that the ordinary skilled person (in this case considered to be a team which included medicinal chemists) would have embarked on a number of conventional strategies, including fractional crystallization, for resolving citalopram, or if those approaches failed, possibly through the use of a suitable chiral precursor. However, in light of the evidence presented, His Honour was not persuaded that the hypothetical skilled person or team would have had an expectation of success using a chiral precursor diol route, and the attack on the method claim on the ground of obviousness failed.

In considering whether the product claims were obvious, His Honour accepted that in 1988 the hypothetical skilled person or team would have recognised the possibility that the therapeutic benefits of the racemate may reside with one enantiomer. However, this was not considered to be sufficient to render the claims obvious. The test applied by His Honour was: “Would the hypothetical addressee……adopt as a matter of routine the aim of obtaining the (+)-enantiomer of citalopram with the expectation that it might produce a useful alternative to, or better drug than, citalopram?”

Bearing in mind his conclusion with regard to the difficulty in obtaining the isolated enantiomers of citalopram, His Honour examined the motivation within the pharmaceutical industry in 1988 for resolving racemic mixtures. He accepted evidence that molecules made as racemates were generally developed as such and that commercial separation of enantiomers was at the time considered to be a “black art” not typically undertaken due to the expense and difficulty of resolution and a general lack of expertise. In particular, citalopram, as the racemate, was known to be a safe and effective antidepressant which was in itself highly selective, and had no adverse side effects, so that there would have been no significant need or desire to obtain the individual isomers. Accordingly, it was held that it would not have been obvious to want to have the separate enantiomers.

On this point, it is worth noting that the inquiry turned on what was the common general knowledge in 1988. The evidence alluded to a shift in emphasis to individual enantiomers during the 1990’s, and accordingly, if the question was to be considered in light of today’s common general knowledge, the conclusion may well be a different one.

His Honour also dismissed an attack on the ground that the claims did not define a new manner of manufacture, finding that the (+)enantiomer was a newly existing entity that has unexpected and surprising properties that could not have been predicted from the properties known for citalopram. Alphapharm’s contentions that the invention was not fully described.and that the claims lacked fair basis and were not clear, were also firmly dismissed.

Alphapharm did however succeed on one ground, that of inutility. On the basis of evidence presented, it was established that the useful minimum dose of escitalopram was 5 mg, with a maximum dose of 40 mg. Claim 5 defined a range of from 0.1 to 100 mg per unit dose and therefore encompassed dosages outside the “useful” range. Given that all the other validity attacks failed, Alphapharm’s success on this point is unlikely to have dealt a great blow to Lundbeck. However, it does serve as a warning to applicants and patentees to carefully consider the dosage ranges recited in unit dosage claims, particularly where the invention lies in the dosage amount, and presents an argument at least for the inclusion of one or more sub-claims defining narrow ranges which clearly only encompass “useful” amounts.

Extension of term

Lundbeck appealed the earlier 2006 decision of the Delegate of the Commissioner directing that the entry in the Register be amended to reflect a term some 18 months shorter than originally recorded, on the basis that the extended term was not correctly calculated on the “first regulatory approval date”. Alphapharm sought rectification of the Register to remove reference to the extension from the Register.

Cipramil (citalopram hydrobromide) was first included in the ARTG on 9 December 1997. Lexapro (escitalopram oxalate) was first included in the ARTG on 16 September 2003. In the present inquiry, the timing of an application for an extension of term required the application to be made within 6 months of the date of commencement of the first inclusion in the ARTG of goods that contain or consist of the relevant pharmaceutical substance (escitilopram). The question before the Court was whether Cipramil qualified as goods that contained the pharmaceutical substance (+)-citalopram, as this would in turn determine which of the two dates was the first regulatory approval date, and ultimately whether the application for extension was timely made. Lundbeck argued that in the present circumstances, the separate (+)-enantiomer is the “pharmaceutical substance per se”, and, pointing to the differing physical, chemical, pharmacological and clinical properties and effects of the (+)-enantiomer and the racemate, contended that it is incorrect to say that Cipramil, the racemate, contains this substance. In dismissing this argument, His Honour considered the per se qualification to apply only to the pharmaceutical substance in respect of determining its disclosure in the complete specification and that thereafter, the question was simply whether the goods, Cipramil, contain the molecule that, from its isolated state, derives the name “(+)-citalopram”. Notwithstanding His Honour’s earlier construction of “(+)citalopram” being “pure” in the sense of something separate from the racemate, he considered that the word “contain” (given its ordinary English meaning) provided a context not present in the patent specification. In affirmatively answering the question His Honour likened the scenario to eggs and milk in a cake – although each ingredient no longer has a separate existence, the cake nevertheless “contains” eggs and milk. Once Cipramil was found to be a good “containing” the (+) enantiomer, the first regulatory approval date was taken to be 9 December 1997 and Lundbeck’s application for extension of term was deemed to have been filed out of time, with the effect that the extension recorded on the Register was expunged.

It does seem somewhat incongruous that for the purpose of considering a patent’s construction and validity, an enantiomer can be considered to be a separate entity but that this element of separateness is then disregarded in an inquiry as to whether that same patent can be extended. Although not argued by Lundbeck, it is interesting to ponder whether the question would have been answered in the negative if unqualified reference to (+)-citalopram was taken to mean “free from the (-)-enantiomer”. Under such a construction, it might be difficult to sustain an argument that Cipramil contains (+)-citalopram free from (-)-citalopram.

Readers familiar with the extension of term provisions will also recognise that in this instance, the relevant deadline for filing the application for extension was not in fact 6 months from 9 December 1997, as stated by His Honour, but 6 months from 27 January 1999, the date when the provisions came into effect. Nevertheless, even on this basis the application for extension would still have been filed out of time.

Protected information

The Therapeutics Goods Act, at s 25A(2), provides that information in respect of applications for the registration of therapeutic goods provided to the Secretary of the Department of Health and Ageing of the Commonwealth of Australia (the Secretary), may under certain circumstances be “protected information”, i.e. the Secretary may not use that information in evaluating other applications for registration. Lundbeck sought a declaration that the information it had provided to the Secretary in respect of various goods containing escitalopram was “protected information” and also an order restraining the Secretary from using that information when evaluating any application for registration of therapeutic goods by any person other than Lundbeck. In 2005 Alphapharm had applied for registration on the ARTG of its generic product containing escitalopram, effectively inviting the Secretary to rely on Lundbeck’s information provided in respect of its registration of Lexapro.

In order to qualify as protected information, a number of criteria must be satisfied, including that when the application was lodged, no other therapeutic goods consisting of or containing, that active component were included in the Register and had not been included any time before then.

In view of the conclusion reached in the extension of term proceeding, His Honour found that prior to the application for registration of Lexapro, the active component (escitalopram) had been contained in therapeutic goods (Cipramil) listed in the Register. Any relevant information was therefore not “protected information” and the applications for a declaration and injunction were dismissed.


Lundbeck (together with its exclusive licensee Lundbeck Australia Pty Ltd) crossclaimed for infringement of claims 1, 3, 5 and 6, alleging that Alphapharm manufactured goods containing escitalopram in order to secure regulatory approval and imported escitalopram made according to the method of claim 6. Alphapharm admitted manufacturing goods containing escitalopram for the purpose of obtaining regulatory approval, but in defence relied on (then) s 78(2) which provided that where an extension of term has been granted, infringement does not occur if the invention is exploited for the purpose of obtaining regulatory approval. Alphapharm also contended that the intermediate used by its supplier was a “bromo-diol” instead of the “cyano-diol” required by Formula II and thus fell outside the scope of claim 6.

Although finding that Alphapharm has infringed the product claims 1, 3 and 5, the Court recognised that a defence lay in s 78(2) (and its eventual replacement, s 119A), for the period commencing after the grant of the extension. However, for the period prior to the granting of the extension the defence did not apply.

The method used by Alphapharm’s supplier utilised a precursor compound where the cyano group of Formula II was replaced by a bromo group. Following the cyclization step, the bromo group was then exchanged for the cyano group (a process accepted as well known and routine) to form escitalopram.

The evidence suggested that notwithstanding some differences between the bromo and cyano groups, the skilled person would have expected the resolution and subsequent cyclization of the racemic bromo-diol to proceed in a similar manner to that of the cyano-diol. On this footing His Honour concluded that there were no contextual differences between employing a bromo-diol and the cyano-diol as claimed and all that Alphapharm’s supplier had done was introduce another step (being the substitution of the bromo for a cyano group). Despite the fact that the process used did not fall within the literal scope of the method claim, infringement was found.

By adopting the approach that a feature of a claim may be substituted with another commonly accepted equivalent as long as the same end result or effect is achieved in the same manner, the line of reasoning taken by His Honour seems to be an espousal of the “doctrine of equivalents”, a feature of United States patent law. Whilst the Australian Courts have long recognised that a non-literal construction is permissible in an infringement inquiry, the conclusion reached here does appear to fly in the face of the well established proposition of “what is not claimed is disclaimed”.

Both Alphapharm and Lundbeck have appealed aspects of the judgment to the Full Federal Court.