By Kurt R. Karst –
Every once in a while FDA gets asked a question along the lines of: “Can my company submit two ANDAs to FDA – from two different subsidiaries – each containing the same bioequivalence data?” . . . or “Can my company sell its bioequivalence data to different companies for ANDA submission purposes?” In each case the answer from FDA has been a flat-out “No.” But why? It all has to do with FDA’s so-called “anti-franchising policy,” which provides that each ANDA for a particular drug product must have its own independent basis for approval. That is, even if a drug product is manufactured using the same formula, same equipment, and at the same facility as an already-approved product, FDA has required the new applicant to perform some level of bioequivalence testing, at a minimum, to support approval.
Although companies have requested FDA to permit them to reference already-approved bioequivalence data, FDA has for decades rejected such requests on the basis of 21 C.F.R. § 314.101(d)(8). This regulation states that FDA will refuse to receive an application if:
The drug product that is the subject of the submission is already covered by an approved application or abbreviated application and the applicant of the submission:
(i) Has an approved application or abbreviated application for the same drug product; or
(ii) Is merely a distributor and/or repackager of the already approved drug product.
Although this regulation does not explicitly state the anti-franchising policy, FDA has interpreted the regulation as such. (By the by, this regulation has also been used as the basis for two subsidiary companies to submit different applications. FDA’s policy has been that the Agency “does not look under the corporate tent.” Thus, the two policies that stem from 21 C.F.R. § 314.101(d)(8) – anti-franchising and corporate tent – have formed the basis for two subsidiary companies to submit different applications, each with its own independent data to support approval.)
FDA’s “anti-franchising policy,” as reflected in the Agency’s interpretation of 21 C.F.R. § 314.101(d)(8), has deep roots. From what we understand, a version of the “anti-franchising policy” policy was discussed in a speech by former FDAer Dr. Peter Rheinstein in Fall 1987. Dr. Rheinstein reportedly explained that one company approached FDA and proposed getting approval for a drug product made in its manufacturing facility and then letting other companies cite its data. FDA rejected this concept.
Not satisfied with a press report about a speech, however, we dug a little deeper and uncovered a letter FDA sent to one company on July 15, 1987. That letter, which we believe informed Dr. Rheinstein’s speech, may be the genesis of FDA’s “anti-franchising policy.” The letter is reproduced below (for posterity and with the company name redacted) and states:
This letter is in response to your April 21, 1987 letter suggesting a different approach for satisfying Agency bioequivalence requirements for abbreviated new drug applications (ANDAs). Your proposed approach involves reference to a drug master file (DMF) for bioequivalence and dissolution studies. I apologize for not responding more promptly but I wanted to give your proposal careful and complete consideration.
I have discussed your letter at length within the Agency and have concluded that each specific drug product must stand on its own, that is, each ANDA must contain information to show that the product is bioequivalent to the listed drug product. In vivo and in vitro tests that you conduct on your own tablets would not satisfy the in vivo bioequivalence requirements for tablets manufactured by someone else. Therefore, reference to your DMF by ANDA applicants would not satisfy the in vivo bioequivalence requirements for the referring ANDA applicant. This is so even though all the tablets involved were made from the same pre-granulated bulk material. A discussion follows.
As you know, the purpose of the Federal Food, Drug, and Cosmetic Act (Act) is to protect the public against danger to human life arising from the use of unsafe and ineffective drug products by assuring that, before any drug product is marketed, it will have been carefully reviewed by FDA experts. With the passage of the 1984 Amendments to the Act and the provisions to allow many additional drug products to be reviewed under abbreviated new drug applications (ANDAs), Congress wanted to assure the continued protection of the public. Congress, therefore, required that to be approved under the amended Act each ANDA must contain information to show that the proposed new drug product is bioequivalent to the drug product it seeks to copy – the listed drug. See 21 USC 355(j)(2)(A)(iv).
In enacting the 1984 Amendments, Congress was cognizant of the FDA regulations then in effect. The term “drug product” as defined it the regulations at 21 CFR 314.3(b) means “a finished dosage form, for example, tablet . . .that contains a drug substance, generally, but not necessarily, in association with one or more other ingredients.” Also, under 21 CFR 314.50 and 314.55 each ANDA is required to contain a description of the manufacturing and packaging procedures and in-process controls for the drug product to ensure, among other things, the bioavailability of the drug product. These sections of the regulations also require the ANDA to contain a full description of the specifications and analytical methods as are necessary to assure the identity, strength, quality and purity of the drug substance and the bioavailability of the drug product made from the drug substance.
As this discussion discloses, a drug product and drug substance are not the same. Assurances about quality or bioavailability, for example, made about the latter cannot automatically or consistently be extrapolated to the former. (In other words, you cannot necessarily extrapolate from the bioavailability or quality of one drug product containing a drug substance to the bioavailability or quality of a second drug product containing the sane drug substance.) The possibility of variations in the manufacturing process (e.g., tableting process) used by each individual drug producer requires, therefore, certain testing on each separate and specific drug product to assure protection of the public. For example, even in those cases where the same ANDA holder (or applicant) changes its manufacturing site and seeks to perform comparative dissolution, studies in lieu of in vivo studies to demonstrate bioequivalence for the products made at different sites, an in vivo test on the finished drug product may be required unless the formulation is virtually identical at both sites and the manufacturing process (including manufacturing procedure and equipment, among others) is identical.
Your proposal would involve reliance on your bioequivalence testing by not just one applicant but by many different ANDA applicants. And, under your proposal, rot only could a different formulation be possible from each applicant, (an ANDA applicant would not be required to add the ingredients you recommend for the proproxyphene/APAP combination, for example), but it is highly unlikely that the tablets you make would be made under the sane conditions and on the identical equipment as those made by each of your customers. It is also likely that each manufacturer might have to modify the formulation of the product to accommodate the manufacturing procedures, equipment and environment at their respective facilities. Thus in time, we could expect “drift” from a once correct formulation to potentially different formulations for each manufacturer.
I have concluded therefore that, where in vivo bioequivalence data arc deemed necessary by the Agency, reference to your DMF by ANDA applicants would not satisfy the bioequivalence requirements for these applicants. Each ANDA applicant is required to include in its own ANDA information to show that its drug product (in its finished dosage form) is bioequivalent to the listed drug. For these reasons, a showing by you that your drug product is bioequivalent to a listed drug cannot satisfy the statutory requirement that other drug products are also bioequivalent to the same listed drug. Each ANDA must stand on its own in this regard to assure the protection of the public health.
Shrikant V. Dighe, Ph.D.
Director, Division of Bioequivalence
Office of Drug Standards
There have been other examples since the one above. For example, from what we understand, FDA required a Syntex generic subsidiary to submit a new bioequivalence study to obtain ANDA approval for Naproxen, even though both the Syntex and the subsidiary products were made at the same plant, by the same personnel, and on the same equipment using the same formula and process. FDA also reportedly required a Merck generic subsidiary to submit a new bioequivalence study to obtain ANDA approval for its generic version of Merck’s DOLOBID (diflunisal) product even though both products were also made identically. In some instances, companies decided to conduct the study and seek ANDA approval. In other instances, companies chose not to proceed with conducting the study and decided to act as a distributor of the brand-name drug product.