When Is Identity of Medical Ingredients Determined for an ANDS?

Apotex Inc v. Canada (Health) 2013 FC 1217 Kane J
            Apo-telmisartan / MICARDIS

The abbreviated new drug submission process allows a generic manufacturer to obtain an NOC for its generic version of a new drug without submitting safety and effectiveness data, if it can establish that its generic version is the “pharmaceutical equivalent” of a Canadian reference product: Food and Drug Regulations C.08.002.1(1)(a). "Pharmaceutical equivalent" means a drug that contains identical amounts of the “identical medicinal ingredients,” but that does not necessarily contain the same non-medicinal ingredients: C.08.001.1. But what does “identical medicinal ingredients” mean, and at what point in the drug production and delivery cycle is identity established?

So, suppose the pure API for the branded product for which an NOC has been granted is a weak base, but it is formulated as a salt form. Can the branded product be used as a reference product if the generic drug is formulated as a pure base? That is, are the salt and the base considered to be the “identical medical ingredient”? Alternatively, once the tablets are swallowed, the salt of the branded product dissolves into the acid and base, and the active moiety in the blood stream is again the base. Can the branded product be used as a reference product on the basis that the medical ingredient is identical in the bloodstream, even if it is not identical in the tablet?

These questions were before the FC in Reddy-Cheminor Inc v Canada (Attorney General), 2003 FCT 542, and both were answered in the negative. Reddy sought to use AstraZeneca’s omeprazole magnesium tablets as the reference product for an ANDS for its omeprazole capsules. AstraZeneca had originally marketed omeprazole in Canada as omeprazole capsules, but in 1995 it submitted a SNDS to change its capsules to omeprazole magnesium tablets. Referring to section C.08.003, Reddy argued that

a SNDS cannot be used to approve a new drug. Since AstraZeneca's SNDS requested approval to change its product from omeprazole capsules to omeprazole magnesium tablets, the TPD must have considered the change, from omeprazole capsules to omeprazole magnesium tablets, not to have been a change in the drug itself. Thus, says Reddy, omeprazole and omeprazole magnesium must be the same medicine and therefore pharmaceutically equivalent. [32]

Layden-Stevenson J called this argument “logical and rational” but she nonetheless rejected it, essentially on the basis that while it is true that an SNDS cannot be used to approve a new drug, there is a difference between not being a new drug, and being an “identical medical product” [34].

Layden-Stevenson J began a subsequent paragraph by stating that “Reddy does not suggest that the drugs in question [omeprazole and omeprazole magnesium] contain identical medicinal ingredients” [62]. I must admit that I do not understand this statement in light of the argument quoted above at [32]. In any event, whether or not Reddy challenged this point, Layden-Stevenson J clearly held that the free base and the salt form are not “identical medicinal ingredients.” The Minister of Health has also consistently maintained this position throughout Reddy-Cheminor, and subsequently confirmed this in its Guidance on Interpretation of “Identical Medicinal Ingredient” 2003-07-03 [4.3].

Reddy also argued that omeprazole and omeprazole magnesium ultimately metabolize to become the same substance, and “[i]t is that substance that has the therapeutic effect and on that basis, it argues that the two products contain the same medicinal ingredient. ANDS.” Layden-Stevenson J rejected this argument as well, saying “I agree with the respondent that the Minister cannot evaluate the specifications of a substance that may be produced following in vivo metabolism. The legislative requirement is to consider, as the medicinal ingredient, the ingredient contained in the drug.” [62]

Thus in Reddy-Cheminor, Layden-Stevenson J held both that a base is not the same medical ingredient as a salt form, and that the comparison must be made of the “ingredient contained in the drug” and not in vivo.

This brings us to Kane J’s Apo-telmisartan decision. Apotex sought to use Boehringer Ingelheim’s MICARDIS as the reference product for an ANDS for its Apo-telmisartan product.The Minister held that MICARDIS could not be used by Apotex as a reference product, because Apo-telmisartan did not contain the “identical medical ingredient” [9]. Kane J held that this conclusion was unreasonable, and she consequently granted the application for judicial review [167].

Unfortunately, the basic facts are not entirely clear to me. Telmisartan is a carboxylic acid [17], and, as best I understand, potassium hydroxide is used in Apotex’s wet granulation tableting process and consequently “the acid-base chemical reaction may take place converting some of the telmisartan into a salt form of telmisartan which would be telmisartan-potassium” [7]. I take this to mean that active ingredient in Apotex’s product is primarily telmisartan in the acid form, but some telmisartan-potassium may be present: see also [17],[20], and especially [18]. However, there are also passages suggesting that Apotex’s product is primarily telmisartan-potassium [9]. With respect to MICARDIS the decision states that “Although Micardis and the generics previously approved are in fact telmisartan-sodium, the only medicinal ingredient listed on the label is telmisartan” [123]. This, and many similar statements (eg [68], [146]), seem to me to mean that MICARDIS is essentially entirely the salt form telmisartan-sodium. So, at [146] “the Minister has acknowledged that Micardis and the generics previously approved convert to telmisartan-sodium [from telmisartan].” The reference to conversion – by SNDS? – indicates to me that the current product is substantially not telmisartan itself. However, there are passages suggesting that MICARDIS is also primarily telmisartan, and that the difference is that sodium hydroxide is used in the wet granulation process, where Apotex uses potassium hydroxide, and MICARDIS is only partially telmisartan-sodium: see eg [100].

Now, if both MICARDIS and Apo-telmisartan were primarily telmisartan, with relatively minor admixture of the potassium or sodium salt, respectively, then there might be an argument to be made that the products are identical because they are both substantially telmisartan. However, Kane J’s reasons do not turn on this argument. That the facts are so unclear as to the precise form of the two products indicates in itself that the point is not central to the reasoning. Rather, Kane J framed the question as being “whether identicality is to be determined at the input stage or at the finished dosage stage” [97], and she accepted Apotex’s argument that “only the input medicinal ingredients can be required to be identical” [113]. Consequently,

[100] While telmisartan-sodium may be formed and be in the finished dosage or tablet form of Micardis and telmisartan-potassium may be in the finished dosage or tablet form of Apo-Telmisartan, neither the sodium or the potassium is the medicinal ingredient; rather telmisartan is the medicinal ingredient in both.

On this reasoning, so long as MICARDIS and Apo-telmisartan are both telmisartan at the input stage, it does not matter that one is telmisartan-sodium and the other telmisartan-potassium at the tablet stage. By the same logic, it would seem to follow that if one product was telmisartan and the other telmisartan-potassium, the medicinal ingredient would also be the same, so long as both were telmisartan “at the input stage.”

I find this very difficult to reconcile with Reddy-Cheminor, which held that the comparison was to be made on the basis of the “ingredient contained in the drug.” Reddy-Cheminor was clear that a omeprazole capsules and omeprazole magnesium tablets are different drugs, and one could not be used as a comparator for the other. Kane J distinguished Reddy-Cheminor on the basis that

[96] In Reddy-Cheminor the input ingredients were different. The Court concluded that the identicality could not be evaluated after the transformation in vivo.

Reddy-Cheminor certainly held that identicality could not be evaluated after the transformation in vivo. But that is not all it held. It also held that the base and the salt of the drug form could not be used as comparators. So the crucial point in Kane J’s distinction is that in Reddy-Cheminor the ingredients were different even at the input stage. The problem is to define “the input stage.” Kane J does not offer a clear definition of this concept, but on the facts, she seems to mean the input to the dosage formulation process, in this case the wet granulation tableting process. So, at [114], Kane J apparently accepts Apotex’s argument distinguishing between “the input ingredient and not the finished dosage form,” by reference to paragraphs C.08.002.1(3) (b) and (c) which refer to the requirement to provide both samples of ingredients of “the new drug” and samples of “the new drug in the dosage form.”

The arguments advanced by Apotex and accepted by Kane J turn primarily on textual considerations, in particular the plain meaning of “ingredient,” which Kane J views as implying inputs [110], and the need for consistency in the use of the term “ingredients” throughout the Regulations [113]-[128].

On the purposive branch of the interpretive exercise, it is not controversial that the purpose of the Act and Regulations is “to bring safe and effective drugs to the Canadian market to advance the nation’s health” [130]. Apotex argued that “the use of excipients (the non-medicinal ingredients) does not affect safety and efficacy” [131]. As I understand it, it is clear that a free base form of a drug and a salt form may be pharmacologically different, so that transformation into a salt form may indeed affect safety and efficacy. This is why salt forms are often patented. So, it seems to me that this argument turns on the point that in Apotex’s process, potassium hydroxide is used for some purpose related to the wet granulation process, and it is not intended to create the salt form. In particular, it seems that the Minister is worried that if excess potassium hydroxide is used in the process, some of that excess may result in the salt conversion [17]. Apotex’s position is that in fact no such conversion occurs [18], so the potassium hydroxide has no effect on safety and efficacy; it truly is an excipient. But Kane J’s reasoning does not turn on this factual point. Her decision implies that the result would be the same if excess potassium hydroxide was added in the tableting process with the intent and result of causing complete conversion to the salt form. In that case, it is doubtful whether the potassium hydroxide could be considered an “excipient” in the ordinary sense of the term, and in any event it would not be true that it has no effect on the safety and efficacy.

The second point made by Apotex on the purposive construction is that “after the screening stage approval of the ANDS further steps in the process would evaluate the safety and efficacy of the drug”[131]. That is, even if MICARDIS can be used as reference product, and even if the telmisartan sodium it contains is pharmacologically different from the telmisartan / telmisartan potassium in Apotex’s product, this is not a concern for health and safety because “the safety and effectiveness would be assessed at the next stages of approval” [140]. As I understand it, it is true that assessing safety and efficacy of a salt form as compared with the pure base does not require as much information as assessing safety and efficacy of an entirely new chemical entity. This is why an SNDS was required when AstraZeneca moved from omeprazole to omeprazole magnesium, as discussed in Reddy-Cheminor. The difficulty is that the SNDS application process requires more information than the ANDS process, and so the Minister does not have as much information at the next stages of approval for an ANDS as it would for an SNDS. This is the purposive reason why Layden-Stevenson J in Reddy-Cheminor held that a “new drug” is different from an “identical medical ingredient:

[34] The specified circumstances requiring the filing of a SNDS include matters such as the plant and equipment to be used in manufacturing, the manufacturing methods, the tests and controls to be applied and (of particular relevance here) the specifications of the ingredients of the drug. The manufacturer cannot market the drug unless the changes have been fully described in a supplement to the original NDS and the Minister is satisfied that the drug remains safe and effective and issues a NOC. This differs from an ANDS that may be submitted provided, among other things, that the drugs being compared contain "identical amounts of the identical medicinal ingredients". No similar provision is prescribed for a SNDS. Approval for an ANDS involves comparison to a Canadian reference product whereas approval for a SNDS involves satisfying the Minister of the safety and efficacy of the drug. There is no requirement that the medicinal ingredients be identical in relation to the latter.

Ultimately, while Reddy-Cheminor and Apo-telmisartan can be formally reconciled, they are very much in tension. Functionally, Reddy-Cheminor turns on the point that an ANDS does not provide sufficient information for safety and efficacy of a salt to be assessed by comparison with a base, while Apo-telmisartan implies that an ANDS does provide sufficient information for such a comparison. This is not to say that Apo-telmisartan is wrongly decided. My difficulties with the purposive analysis in Apo-telmisartan are hardly conclusive even on that point, and the textual and contextual arguments which Kane J found compelling are undoubtedly crucial aspects of statutory interpretation. But it does seem to me that there is a tension between the purposive and textual considerations, as well as between the cases. If this case is appealed, I trust that the FCA will resolve this tension, one way or the other.