Dockets ManagementBranch (HPA-305) U.S. Food and Drug Administration 5630 Fishers Lane Room 1061 Rockville, Maryland 20852 Re: Docket No. 2004P-0506: Comments of Noven Pharmaceutical.Inc. on ALZA Cornoration Citizen Petition ReouestinaAction Regarding Generic Fentanvl Twnsdmal

These comments are submitted by Noven Pharmaceuticals,Inc. (“Noven”) in opposition to the above Citizen Petition (the “Petition”). The Petition was filed by ALZA Corporation (“‘ ALZA”), the manufacturer of the patch technology used in the branded fentanyl transdermal product, Duragesic@. In filing the Petition, ALZA seeksto have its proverbial cake and eat it, too. ALZA first affirms that it “supports FDA approval of generic fentanyl transdermalproducts.“’ Scientific integrity requires no less. Yet, in virtually the samebreath -- and based on conjecture, incorrect assumptionsand flawed studies -- ALZA asks FDA to require manufacturersof generic fentanyl transdermalproducts “to develop and implement” Risk Minimization Programs(“RMPs”) and to classify such products as different dosageforms from the brandedproduct, DuragesicQ rather than as pharmaceuticalequivalents.

’ Petition at 1.

2QowP-OS -

Ob 11960 SW 144th Street V Miami, Florida 33186 305-253-5099 T Fax: 305-251-1887


At the outset,it is noteworthy that, true to its word, ALZA’ “Actions Requested” not s do include a requestthat FDA require that RMPs be in place prior to ANDA approval.* Again, scientific integrity demandsno less. Holding ALZA to its word, FDA should proceedto ANDA approval in time for the January23,2005 launch of the generic versions,and considerthe need for RMPs for Duragesic@and generic fentanyl transdermalproducts at a later date, if at all. However, to the extent ALZA is suggestingthat FDA should require RMPs as a condition of approval for generic formulations of fentanyl transdermal,there is neither a scientific nor legal basis for such a requirement.3In fact, it is quite remarkablethat ALZA would have the temerity to petition FDA for an RMP for its genericcompetitorswhen there is no RMP in place for its brandedproduct DuragesicO,especiallyin light of the reported casesof abuseof Duragesica tied to the reservoir design of its delivery system. Requiring RMPs for only generic versionswould be unprecedented quite anomalous. and ALZA also requeststhat FDA reclassify solid statematrix patchesas a different dosage form from the reservoir patch. By this ruse, ALZA seeksto delay approval of the ANDAs and to incapacitatethe generic manufacturersfrom marketing their products and competingon a level playing field with ALZA’ brandedproduct, Duragesic@.However, this request,as well, is both s scientifically and legally unsustainable. As discussedherein:

ALZA doesnot even contestthat solid statematrix fentanyl generic transdermalproducts are safeunder the conditions of use prescribed, recommendedand suggested the proposedlabeling, the ultimate legal in standardthat guidesFDA’ drug approvalprocess. s

action is that FDA requireRh4Psfor genericfentanyltrausdermal products,but does ’ Petition at 1. The requested not specify that thoseRMPs would needto be developed implemented and prior to ANDA approval. 3 Petitionersmay be statinggiugerly that FNPs shouldbe a conditionof approvalin the last sentence the Petition, of wherethey remarkthat “Product specific risk minimization programsareneeded supportthe introductionof to products,including fentanylmatrix productsthat may presenta greaterpotentialfor diversionand abuseiu the US environment.”Petition at 9.



ALZA’ position is severelyundercutby the facts that (1) it believesANDAs s for fentanyl patchesshouldbe approvedand (2) ALZA is taking off the market in Europethe fentanyl reservoirdesign(usedin the Duragesic@ patch in the U.S.) and replacingit with the solid statematrix designthat ALZA attacksin this Petition and that hasbeenapprovedin Europeancountries. The scienceadvanced the Petitionercannotwithstandscrutiny. Put simply, by the solid statematrix systemdoesnot posethe risk ALZA purportsto attribute to it. FDA has alreadyruled’ different mechanisms delivery do not place that of productsin different dosage form categories.The ruling was basedon a detailedanalysisof the legislativehistory of the Food, Drug and CosmeticAct (‘ FDCA”) and soundpolicy considerations light of the Act’ purposes.The in s court upheldFDA’ ruling. Thus, this is an unproductiveargumentfor ALZA s to rely upon.



In any event,the designof Noven’ genericfentanyl transdermal s patch is as safeas is Duragesic’ underthe conditionsof userecommended the sG9 in proposedlabeling -- and, indeed,lesssubjectthan is Duragesic@ m isusein to the form of abuseand diversion.

In sum, ALZA’ Petition is nothing more than a thinly-veiled attemptto extendits s s monopolypower of its DuragesicBproductbeyondthe time Congress allotted to it. ALZA’ double standard one for its product, anotherfor potential competitors-- is problematicfrom -s both legal and scientific perspectives.It also collides with Congress’goal of giving innovators exclusivity to reap the benefitsof monopolypricing for a certainperiod of time, and then permitting the public to reapthe benefitsof competition. SinceALZA’ nakedattemptto extend s its monopoly and stifle genericcompetitioncan withstandneither scientific nor legal scrutiny, ALZA’ Petition shouldbe denied. s I. INTEREST OF NOVEN PHARMA CEUTICALS. INC. Noven is a leadingU.S. manufacturer prescriptiontransdermal of patches,including the type of “matrix” patch addressed ALZA in its Citizen Petition. Noven haspartneredwith by Endo Pharmaceuticals to bring to market a genericcontrolled-release Inc. fatany transdermal patchusing Noven’ matrix transdermal s system. Noven filed an ANDA for fentanyl transdermal 3

systemon July 30, 2003.4 FDA acceptedthe ANDA for filing on October 1,2003. ALZA’ s Citizen Petition seeksimproperly to delay FDA approval of Noven’ ANDA and to place s burdenson Noven’ and Endo’ ability to market their generictransdermalproduct on a level s s playing field in competition with ALZA’ Duragesic@.AL,ZA’ characterizations the s s of delivery systemsused in generic fentanyl transdermalproducts like Noven’ are inaccurateand s misleading. For thesereasons,Noven has an interest in the subjectmatter of the Petition within the meaning of 2 1 CFR $10.30(d),and thus respectfully submitsthis response.Noven requests that the Agency deny the action requestedby ALZA. On December 10,2004, Noven submittedcommentson an earlier petition, in Docket 2004P-0472,which also involves FDA approval of generic formulations of fentanyl transdermal (“Brookoff Petition”). A copy of Noven’ submissionis appended s hereto as Attachment 1. Many of the argumentsmade in the Brookoff Petition are echoedin the ALZA Petition that is the subject of the instant Comments. Thus, Noven’ commentson the Brookoff Petition addressin s detail the numerousfallacies in the scientific and legal contentionsthat purport to support the overlapping argumentsin the two petitions. We will not expandthe Agency’ burden by repeatinghere the points we made in detail s in our responseto the earlier Petition. Rather,Noven incorporatesby referenceits Commentson the Brookoff Petition in Docket 2004P-0472and will focus here on argumentsraised in the ALZA Petition that either were not presentedin the earlier petition or were offered with a somewhatdifferent thrust.

’ ANDA 76-804.



transdermalfentanyl patchesare more likely to be abusedthan is its own product, ALZA requeststhat FDA require that all applicantsfor generictransdermalfentanyl products develop and implement RMps approvedby FDA. However, Congresshas clearly defined and limited the criteria that FDA can require in an ANDA. FDA has previously recognizedand respected those limitations in connectionwith establishmentand implementationof RMPs for genericproducts and has never required an RMP as a condition of ANDA approval. ALZA doesnot provide any legal authority for FDA to require an FMP here. Thus, FDA cannotmake final approval of any ANDA for a fentanyl transdermalmatrix patch contingentupon the submissionof an RMP. We have provided in detail the uncontrovertedlegal authority that supportsthis position in our Commentson the Brookoff Petition.’ We will not repeatthat discussionhere, but rather incorporateit by reference. III. ALZA HAS FAILED TO PRESENT ANY COMPETENT SCIENTIFIC 3 E ‘ NCE 0 FENT ANYL TRANSDERMAL MATRIX SYSTEMS THAT IS SUFFICIENT TO NECE$SITATE RMPS FOR GENERIC FORMULATIONS. While Noven recognizesthe utility of RMPs in certain circumstances, ALZA misstates the potential for abuseof matrix products such as Noven’ in suggestingthat RMPs for generic s fentanyl formulations are necessary.Indeed,the “possible”abuseALZA suggests basedon is mere conjectureand is without any sustainablescientific support,especiallyconsideringthe nature of the solid statematrix design6 In advancingits hypothesis,ALZA downplaysthe

5 See Attachment 1 at 2 l-23. ‘ See id. at 6-7, 11-14.

capacity for abusethat has actually been reportedwith Duragesic@ usesfalse comparisonsto and draw improper conclusionsabout the relative likelihood of diversion of the matrix formulation. As Noven discussedin detail in its Commentson the Brookoff Petition, there is no basis for an argumentthat Noven’ matrix design is subjectto greaterpotential abusethan is Duragesic@; s in fact, Noven’ product will likely be less attractiveto abusers s than is Duragesic@.’ a result, As there is no need for Noven to implement an FWP for its fentanyl transdexmal patch, especially consideringthe fact that, to our knowledge,ALZA has not implemented,or even proposed,an RMIP for Duragesic@. ALZA postulatestwo incorrect reasonswhy fentanyl matrix formulations might be subjectto greaterabusethan the DuragesicQsystem: (1) matrix patchescan be cut into pieces, creating “unit doses,”and (2) when soakedin various solvents,there is a greaterpercentage yield of fentanyl from matrix systemsthan from the Duragesic@.ALZA attemptsto supportits conjecturewith a flawed study that it commissioned. Neither the argumentsnor the data are persuasive. A. ALZA’ Claims of Potential Abnse Based on the De&n of Matrix Svstems s are Flawed.

ALZA first arguesthat the fentanyl transdermalmatrix systemsare likely to be abused basedon two design elements. Both of ALZA’ theories-- that either cutting matrix patchesor s attemptingto extract fentanyl through soaking of matrix patcheswill facilitate abuse-- are wrong.

‘ Seeid. at 11-21.


Cutting the Solid StateMatrix Patch into Smaller Units Does Not Impact the Rate at which Fentanvl is Delivered.

The fact that the matrix patch can be cut into piecesis immaterial. As discussedin great detail in Attachment 1, Noven’ submissionin Docket 2004P-0472,*the controlled-release s of fentanyl from a solid statematrix patch results from the combination on a molecular level of fentanyl with two different pressure-sensitive adhesives.No amount of cutting can compromise the releaseof drug from such a drug-in-adhesivesystem,as releaseoccurs through the competing powers of diffusion, acting to push the drug out, and the attraction of the drug for the adhesives, acting to hold it in the system. As a result, cutting the Noven patch into “tmit doses”for sublingual or buccal exposurewill not result in the user deriving a euphoric dose. Although the rate of drug absorptionmight be expectedto accelerate somewhatwhen placed in the mouth, the fact is that the amount of drug releasedfrom the solid statematrix patch would be limited by the controlled rate of diffusion of the drug from the adhesivepolymers. Thus, the net result from cutting Noven’ matrix systemwould be the creation of multiple smaller slow-release s units, which would have no addedbenefit to abusersand no increasedpotential for abuseor diversion. 2. Usable Dosesof Fentanyl Cannotbe Easily ExtractedFrom Matrix Patches.

ALZA’ suggestionthat fentanyl can be more easily extractedfrom matrix systemsfor s smoking or injection is equally flawed. Soakingthe patch in a substance such as methanol,a metabolic poison, for severalhours is neither a quick nor easyway for an abuserto gain access to the opioid. In fact, after soaking the patch, the user would need to undertakefurther stepsto extract the drug from the noxious adhesives mixed with it in solution, as well as removing the methanol itself.

*See Attachment 1 at 6-7, 13-14.

ALZA’ argumentthat soakingthe DuragesicQpatch results in lower yield of fkntanyl s than soaking a matrix patch is misplaced. ALZA ignores the fact that virtually the entire concentrateddose of fentanyl containedin the reservoir of the Duragesic@ patch can be obtained merely by cutting open the reservoir, which can be done quickly and easily. In comparison,any fentanyl obtained from soaking a matrix patch would be mixed with chemicalsthat would s grossly degradethe purity of the drug, as noted above. Thus, in contrastto ALZA’ “Room TemperatureSoak”chart measuringthe yield of fentanyl in solution after severalhours of soaking,ga user can obtain a virtually undiluted yield of abusablefentanyl in secondsfrom the dose of fentanyl from Duragesic@ patch.lo Certainly, the ability to manipulatea concentrated the gel in the DuragesicQreservoir for immediateuse (without further needto extract the drug from solution) would be more attractive to an abuserthan attemptingto extract fentanyl from the s matrix patch. The latter, by ALZA’ own admission,can only be done after severalhours of extraction in chemical solvents,followed by fkther time and effort to extract the drug from the adhesivemixture for purposesof injection or smoking.” In sum, there is no reliable evidencethat the design characteristics the Noven matrix of patch will lead to the potential abusethat ALZA cites as the basis for its demandfor RMPs for generic formulations of fentanyl transdermalproducts.

9 Petition at 5. lo This dose,contraryto ALZA’ implication, canbe split into multiple, usabledosesmerely by applyingthe s fentanylgel to gauze. SeeAttachment1 at 14. Gtherusershavedeveloped more creativeways to portion this otherwisepotentially lethal dosein orderto abusethe product. Id. at 15, x1.32. ” The fact that ALZA also soakedits own matrix patch in non-toxicliquids, such as nun and vodka,and obtained in s on significant yields after severalhoursis also immaterial. First, as discussed Noven’ Comments the Brookoff Petition (Attachment1 at 13), fentanylhaslower bioavailability and a slowerrate of absorptionwhenabsorbed of throughthe GI tract, limitiug the euphoriceffect of driuking the fentanylin solution. Moreover,regardless the toxicity of the extractingagentitself, ALZA’ argument s againignoresthe fact that the silicone and acrylic adhesives utilized in Noven’ patchwould be co-extracted thereafter s and needto be removed.



The Studv ALZA Cites in Support of its Argument Reeardiw Potential Abuse is Unreliable.

In an attachmentto its Petition, ALZA presentsthe results of an unpublishedstudy it commissionedthat ALZA improperly contendsconstitutesscientific support for the notion that a fentanyl transdermalmatrix systemwill be more “attractive”to abusersthan the Duragesic@ system. While ALZA describesthe study as “valid and reliable,”there are substantialquestions about the internal and externalreliability of both the study and the scaleusedto measure potential abuse. With respectto the internal reliability of the study, the researchers themselves note that “nearly a quarter of both the developmentaland confirmation samplesclaimed experiencewith the fentanyl matrix patch, which is not available.“‘ The only explanation * provided is the possibility that the study participantsdid not take the task seriously,were not entirely truthful, had a tendency“towards braggadocio”or misunderstoodthe research materials.l3 Frankly, both the problem and the explanationare mind-boggling. First, nearly 25 percent of the study participants claimed experiencewith a nroduct that doesnot exist! Then, on top of that, ALZA offers the even more amazingexcusethat the study participantsdid not take the whole thing seriouslv or were lvinn. And, this is the data ALZA would have the U.S. Food and Drug Administration rely upon! This is hardly an endorsement the study’ credibility, or, of s for that matter, ALZA’ s. The study is also externally unreliable. First, the number of participantsin the study -- 40 total, with 5 subjectsor fewer listed as casualabusers-- is extremely small. Second,the researchers themselvesacknowledgepotential bias in the ratings provided by the participants,

I2 Petition Exhibit 2 at 78 (emphasis added). I3 Id.

becausethe determinationsof attractiveness were basedon information provided by the researchers cardsrather than on independentexperiences the participants.14The on by “information” ALZA provided to the participantshas not beenprovided with the Petition. If the information provided to the participantsmisstatedthe nature and potential for abuseof Noven’ s patch in any way close to ALZA’ mischaracterization matrix patchesin its Petition, the s of participants’ responses essentiallyworthless. Moreover, since the participantshad no actual are experiencewith the difficulty of obtaining fentanyl from a matrix system,responses that that systemwas hypothetically attractive are not accurateindicators of the real potential for abuseof such products. Third, drug abuseresearchers have never reachedagreementas to a “Gold Standard”for measuringopioid attractiveness there is no reasonto believe that the standard and proposedin the study will becomesuch a “Gold Standard.”Absent such a standard,and in light of the fact that this ALZA-commissioned study has not beenpeer-reviewed,there is nothing to commendthe method utilized -- agreementbetweenthe participantsas to attractiveness as a -valid measureof potential for abuse. Even the researchers l5 point out that “attractiveness.. is . only a componentof predicting whether an opioid product will be abused.“16Thus, the study cited by ALZA provides no reliable support for a claim that RMPs are necessary generic for fentanyl products. C. ALZA’ ComDarison to OxvConth@ is Disineennous and Inanropriate. s

ALZA not only relies on this flawed study to suggestthat generic fentanyl is subjectto heightenedabuse,but suggeststhat in light of the abusepotential, “[t&is situation is somewhat

I4 Id. at 79.

Is The researchers understate point in noting that “the demarcation this betweenreliability analyses validity and analyses less clear.” Id. at 30. is
I6 Id. at 79.


increasingprevalenceof drug abusein Europe and Europol’ increasedfocus on drug abuse s throughout Europe. Baseline EMCDDA data from 1999 for opioids alone reveal that the potential for abusein Europe is not as different as ALZA claims. For example,in the United Kingdom, 18% of 15 to 16 year olds surveyedin 1999 perceivedheroin as “very or fairly easy” to obtain; that percentagewas similar to most other counties in Eu.rope.20 Moreover, under Europeangovernment-sponsored AIDS preventionprogramsin which syringesare made availableto drug abusers,27 million syringeswere distributed in the United Kingdom in 1997; in 1999 almost three million syringeswere distributed in Portugal; and nearly a million syringes were distributed that sameyear in Austria.2’ Basedon thesedata, ALZA’ bald contentionthat “the environmentregarding s prescription drug abusein Europe differs from that in the US and a significant problem with matrix systemswould not be anticipated”is unsupportable.ALZA doesnot cite to any expert opinion (though it mysteriously mentions such an opinion without any citation) and fails to elaborateon the factors presentin Europe that somehowreducethe risk of potential abuseof an opioid. Moreover, one published study has expresslydescribedinstancesof fentanyl abuse(in non-transdermalformulations) in Europe.22In light of the fact that abusersin Europe have experiencewith fentanyl, the lack of reported abusein Europe of the fentanyl matrix system marketedby ALZA’ affiliate arguesagainstALZA’ dire predictions of abusefrom the s s marketing of fentanyl matrix formulations in the United States.

20EMCDDA 8z Europol,European Union Strategy on Drugs 2000-2004 15,21,24,31,40,46,55 (Draft, Snapshot 1999-2004) (Version 27/07/04),available at
*’ Id. at 44,46,57.

22KronstrandR, Druid H, HolmgrenP, Rajs J. A clusterof fentanyl-related deathsamongdrug addictsin Sweden.
Forensic Sci Int 1997;88: 185-195.


In sum, there is no scientific basis for ALZA’ speculationsabout abuseof Noven’ s s matrix systemor its suggestionthat Noven’ matrix delivery systemis more likely to result in s abuseand diversion of fentanyl than the current reservoir systemused in DuragesicB. IV. RE IS NO BASIS FOR TREATING NOVEN’ FENTANYL S TRANbDERMAL MATRIX SYSTEM AS A DIFFERENT DOSAGE FORM DIJRAGESIC@. The FederalFood, Drug and CosmeticsAct requiresparallels in conditions of use, active ingredients,route of administration,dosageform, strength,bioequivalencyand labeling to ensure that the generic product, like the innovator product, is safe and effective under the conditions of use prescribed,recommendedor suggested the proposedlabeling of the drug.23Under this in rubric, FDA has previously determined(i) that a patch is a single dosageform and (ii) that different mechanismsof delivery betweena brandedand generic drug are not a lawful basis for characterizingthe drugs as having different dosageforms. ALZA’ suggestionhere that FDA s should arbitrarily deem the transdermalpatch utilized in Noven’ product to be a different s dosageform from ALZA’ transderrnalpatch is inconsistentwith the law, with FDA’ prior s s position and has no valid scientific basis. k ALZA’ Reauest that FDA Designate Trawl s mal Matrix Patches and Transdermal Reservoir Patehesas Different zaze Forms is Inconsistent with Federal Lwal Precedent and Policv.

ALZA arguesthat the lack of a rate-controlling membranein a matrix product might Sect the rate of fentanyl delivery and, as a result, such products should be classified as a different dosageform from reservoir patches.24 Under the Hatch-Waxmanamendments the to FDCA, FDA must determinewhether certain enumerated characteristics an ANDA applicant’ of s product are parallel to that of the referencelisted drug. In addition to showing sameness in
~321 U.S.C. 4 355(j)(2)(A); 21 U.S.C. 0 355(d). u Petition at 7.


similar to FDA’ approval of generic versionsof OxyContin@.“” Yet, in anothershow of trying s to walk a tightrope, only two pagesearlier ALZA admits that abuseof DuragesicB and fentanyl s is significantly lower than reportedcasesof abuseof OxyContin@. As ALZA’ own self-serving statements confirm, the suggestionthat fentanyl is subjectto the type of abusethat has been reportedwith OxyContin@is absurd. Moreover, in the OxyContin@situation, the branded manufacturerhad begun working with FDA on implementing a substantialF&P almost a year before any generic formulation even obtainedtentative approval.‘ The generic oxycodone * manufacturersvohmtarily agreedto implement RMPs for their generic oxycodoneformulations prior to marketing. Those facts are not similar to the current situation; they are in stark contrastto it. In the situation at issue,the risk of abuseALZA attributesto solid-statematrix formulations is speculative,there is no evidencethat the generic formulation will be abusedat either a greateror even equal rate as the brandedproduct, and ALZA has not implementedor proposed-- and FDA has not suggested need for -- an RMP for its brandedproduct DuragesicB. the

ALZA’ sE rien e in Euro S * Svstems are not More Likek to be Abused.

Finally, ALZA assertsthat its experiencemarketing fentanyl transdermalmatrix systems in Europe is irrelevant becauseopioid abuseessentiallyis not a problem in Europe as it is in the United States.lg This contention flies in the face of the EuropeanUnion’ establishment the s of EuropeanMonitoring Centre for Drugs and Drug Addiction (EMCDDA) in responseto an

” Petition at 6. ‘ PurduePharmaL.P. had beenworking with FDA on a proposed * Rh4Pas early as August 2001. SeePetition for Stay of Action, Docket No. 04P-0006, PSA-1 at 5 n.9 (filed Jan.7,2004). FDA tentativelyapproved first the ANDA for a genericformulation of extended-release oxycodone hydrochloride(filed by Endo Pharmaceuticals Inc.) on July 3 1,2002. l9 Petition at 3.


conditions of use, active ingredients,route of administration,dosageform, strengthand labeling, the ANDA applicant must demonstrate bioequivalencybetweenits proposedgenericproduct and mechanism-- bioequivalencethe brandedproduct.25As a result, the statutecontainsa separate - by which FDA ensuresthat the rate and amount of drug deliveredby a generic applicant’ s product are equivalentto thoseprovided by the referencelisted drug.26The dosageform requirementdoesnot and neednot encompass and amount of drug delivered, as an ANDA rate applicant will be required to demonstrate equivalentrate and amount of delivery under the the bioequivalencyprovision. FDA has previously ruled that a drug’ dosageform should not be basedon the product’ s s releasemechanism,since rate of releaseis already adequatelyregulatedby the bioequivalency requirement. In its responseto a Citizen Petition requestingthat FDA redefine dosageforms with varying releasemechanismsas distinct dosageforms, FDA explainedthat in implementing the FDCA’ provisions through regulationsit had found “no scientific basis for distinguishing s dosageforms on the basis of releasemechanisms,” that the ‘ and bioequivalency standardsassure the therapeuticequivalenceof any pharmaceuticallyequivalentextended-release product.“27 FDA noted that the term “dosageform” is not defined in the Act, and that when Congress amendedthe FDCA in 1984 to establishthe abbreviatedapprovalprocessfor generic drugs it neither required generic drug products to be identical in all respectsto innovator productsnor overturnedFDA’ establishedsystemof dosageforms, under which no distinction in dosage s form was basedon the releasemechanismof the products.28
” 21 U.S.C. 0 355(j)(2)(A)(i) - (iv). 26See21 C.F.R. 0 320.1(e)(definition of “bioequivalence” FDA’ regulations). in s 27FDA Response Citizen Petition by Pfizer Inc., Docket No. 93P-0421 5,11 (Aug. 12,1997). (“Nifedipine to at Petition Ruling”).
2aId. at 3, 5.


In its ruling on the nifedipine Petition, FDA reviewed its current dosageform classificationsand specifically pointed to extended-release patchesas a type of dosageform that included productswith varying releasemechanisms.29 FDA explainedthat, while theseproducts might vary in “the way they ‘ house’ drug, the ‘ the reservoir’of drug, and the size of the s patches,” they were all the samedosageform.3o Thus, the fact that Noven’ fentanyl transdermal matrix patch usesa delivery mechanismthat differs from that used in the Duragesic@ transdermalfentanyl patch doesnot result in the productshaving different dosageforms under the FDCA. Despite the technologicaldifferencesbetweenthe reservoir and solid statematrix mechanisms, both the brandedand generic formulations of fentanyl are delivered transdermally through a patch. FDA makesno distinction betweenpatch technologieswhen considering whether the dosageform of an ANDA product is parallel to that of the innovator product.3’ When FDA’ ruling was challenged,the court agreedwith FDA’ interpretationand held s s that, as long as “a generic drug falls within the samedosageform classification (as defined by the OrangeBook) as the pioneer drug, it will meet the thresholddosageform ‘ sameness’ requirement”in the Hatch-Waxmanamendments.32 court further relied on FDA’ ruling and The statedthat “a drug’ dosageform is not basedon its releasemechanism,but on its physical s appearance the way the drug is administered.“33 thesereasons,there is no principled and For

291d. at5 soId.

” Food andDrug Administration,Centerfor Drug EvaluationandResearch, AnnrovedDrumProdue@ with Theraneutic Euuivalents,(“The OrangeBook”) 24th Ed., “AppendixzUniform Terms”(listing Dosageforms of drug products).
32Pfizer Inc. v. Shalala, 1 F. Supp.2d 38,46 (D.D.C. 1998),af’ inpart and rev’ inpart, 182F.3d 975 (D.C. Cir. d d

1999)(holding that only the dosage form must be identicalto that of the pioneerdrug; the release mechanism the for genericproduct,which is considered be part of the compositionor formulationof a drug, doesnot needto be the to sameas that of the pioneerdrug in orderto satisfy the ANDA requirements.).
33Pf2er Inc. v. Shalala, 1 F. Supp.2d at 44.


basis for distinguishing betweenthe dosageforms of transdermalmatrix patchesand reservoir patches. As discussedin our commentson the Brookoff Petition, Congressenactedthe HatchWaxman amendments the FDCA principally to createa more expeditiousand lesscostly to regulatory processfor FDA pre-market approval of genericversionsof previously approved brand-namedrugs after the brandedmanufacturers’ generousperiods of exclusivity had expired.34This processenablesgeneric formulations “to be marketedmore cheaply and

FDA’ position on the appropriatebasis for distinguishingbetweendosageforms is s consistentnot only with the statutory criteria set forth in the FDCA and its legislative history for evaluation of an ANDA, but also with the goal of encouraginggeneric competition and innovation.36In its ruling on the nifedipine Petition, FDA explainedwhy its decision constituted good policy, stating that the dosageform categoriesestablishedin the OrangeBook: have effectively servedthe public, the Agency, and the industry. The categories are useful in that they are sufficiently differentiated to make a reasonable distinction basedon dosageform, which includesthe appearance the drug. of However, the categoriesare also useful in that they are not so narrow as to be virtually product-specific. As a result, thesecategorieshave allowed the FDA to make threshold determinationsthat productshave the samedosageform while encouragingmanufacturersto develop innovative releasetechnologiesand allowing the public the benefit of safe and effective generic drug products.YJ7

3’H.R. Rep.No. 98-857,pt. 1, at 14 (1984)reprintedin 1984U.S.C.C.A.N.2647. 35Eli Lilly & Co. v. Medtronic, Inc., 496 U.S. 661,676 (1990). Congress recentlyamended theseprovisions,in largepart to curb abuses pioneersseekingto extendthis monopolybeyondCongress’intent. SeeMedicare by s PrescriptionDrug Improvementand ModernizationAct of 2003,Pub. L. No.lOS-173(117 Stat.2066) $1 1lOl1103,1111,1117(2003). 36In its nifedipine Petition ruling, FDA explainedthat the Petitioner’ request FDA redefineits dosage s that form classificationsystembasedon the releasemechauisms the productswould “underminethe purposeof the Act’ of s genericdrug provisions.” Nifedipine Petition Ruling at 6. 37Nifedipine Petition Ruling at 6-7; seealso Pfizer Inc. v. Shalala, 1 F. Supp.2d at 44 (quotingFDA’ decisionwith s approval).


In fact, fentanyl transdermalmatrix systemsare an exampleof genericmanufacturers developing an innovative releasetechnologythat presentsnumerousimprovementsover the brandedproduct and may actually result in less abuseand diversion of the drug than is experiencedwith the referencelisted drug product.38Furthermore,as FDA itself has noted, a dosageform classification systembasedon releasemechanismwould severelylimit the number of generic drug products that pharmacistsand physicianscould substitutefor referencelisted drugs, and this %ould undermine Congress’ intent to make low cost generic drugs availableto
,939 consumers.

Thus, FDA should maintain its position regardingthe proper meansfor distinguishing betweendosageforms. To do otherwisewould result in reevaluationof the limits of a drug’ s dosageform eachtime a brandedproduct’ exclusivity were about to expire and would create s substantialimpedimentsto generic competition, as genericmanufacturers would have no way to predict whether their products in developmentwould qualify as the samedosageform as the brandedproduct with which they seekto compete. B. ALZA’ Palported Scientific Grounds for Class&inn Matrix and Reservoir s Patches as Dif’ ferent DosageForms Are InsnmortabIe.

ALZA attemptsto side-stepthe lack of any legal basis for determiningthat the products are different dosageforms by suggestingthat two external conditions might impact the rate of releaseof fentanyl from a matrix patch and that, as a result, Noven’ generic formulation should s be classified as a different dosageform for scientific reasons. None of the argumentsALZA presents,however, is capableof withstanding scientific scrutiny.
38SeeAttachment1 at 14-18. ALZA hasrecognized advantages the inherentto a fentanyltransdermal matrix systemand the removalby ALZA’ European s affiliate of its fentanyltransdermal reservoirpatchin the marketsin which it has introduceda matrix patch leadsto the conclusionthat ALZA itself believesthat matrix systems represent improvementover a fentanyltransdennal an reservoirpatch. Theseadvantages includeless likelihood of fentanylleakage.SeePetition at 3. 39Nifedipine Petition Ruling at 7.



Noven’ Fentanyl TransdennalMatrix SystemDoes Not Presenta s HeightenedRisk For PatientsWith Strinned Skin.

First, ALZA suggests the rate of releasethrough a matrix systemmight be increased that when usedby patientswith compromisedskin. ALZA makesno independentargumentto support this supposition,insteadrelying on a separately-filedCitizen Petition (Docket No. 2004P-0340). In its responseto that Petition, Mylan TechnologiesInc. (“Mylan”), another applicant seekingapproval to market a generic fentanyl transdermalpatch, decisively rebutted the argumentsin the Petition cited by ALZA. That Petition drew incorrect conclusionsfrom the two studieson which it relied, including the study cited in ALZA’ petition, and Mylan s effectively noted the flaws in the Petitioner’ arguments.4o the reasonsset forth in Mylan’ s For s responsein Docket 2004P-0340,ALZA’ argumentthat the rate of fentanyl delivery might be s higher in matrix userswith compromisedskin is incorrect and doesnot warranf classifying matrix systemsas different dosageforms. As noted by Mylan, use of transdermalfentanyl systemson compromisedskin is contraindicatedin the product labeling for Du.ragesic@ in the proposedlabeling for the and generic transdermalmatrix systems. Specifically, the Duragesic@ labeling statesthat the product “should be applied to non-initated and non-irradiatedskin on a flat surfacesuch as chest,back, flank or upper arm. . . . Hair at the application site shouldbe clipped (not shaved)prior to system application. If the site of DUFLAGESIC@ application must be cleansedprior to application of the system,do so with clear water. Do not use soaps,oils, lotions, alcohol, or any other agents

40Mylan demonstrated that (1) the observation that stripped skin is more permeable than intact skin is well-known and was no doubt contemplated by FDA in review of ANDAs for matrix systems; (2) the studies cited in the Petition do not contain data supporting the Petitioner’ conclusions thatapplication transdermal matrix systems to stripped s of skin can result in toxic fentanyl blood levels; (3) the study of the Cygnus patch that was relied upon by the Petitioner involved a product that is not a generic equivalent to Duragesic@ and its data are therefore not applicable to generic transdermal fentanyl products; and (4) the use of the patch on stripped skin is contraindicated in the labeling for Duragesic and its generic equivalents.


that might irritate the skin or alter its characteristics.Allow the skin to dry completely prior to systemapplication.‘ In addition, the patient information distributed with Duragesic@ “’ instructs patients as follows: “Do not put the DURAGESICB patch on skin that is very oily, burned, broken out, cut, irritated, or damagedin any way,‘ The labeling and patient information for *2 Noven’ generic fentanyl transdermalproduct will contain the samewarnings and instructions. s Furthermore,AIZA has presentedno data to indicate that mere removal of an adhesive bandagewould result in skin so compromisedthat rapid absorptionof fentanyl would result. In fact, the only in viva quantitative study performed to examinethis issue showedlittle changein the amount of drug that was absorbed(penetrationof hydrocortisoneon the forearm increased from 1% to only 3%), despitethe fact that the study involved stripping of the stratumcomeum to the glistening layer, where water loss is so high that the skin appearswet.43 Thus, it cannotbe assumed,as ALZA would have the Agency do, that routine incidental application of adhesive tape will have a significant effect on the delivery of any transdermalmedication. 2. Noven’ Fentanyl TransdermalMatrix SystemSimilarly Does Not Present s a HeightenedRisk With Resnectto Heat.

ALZA then arguesthat becauseheat will impact the rate of releaseof fentanyl from matrix patches,thoseproducts should be classified as a different dosageform. This argument fails because,as ALZA acknowledges, “[i]t is well established the application of heat to a that fentanyl transdermalsystemenhances delivery.‘ This statementis true regardlessof the design * of the delivery system. That is why the labeling for DuragesicB -- and Noven’ proposed s
41SeePhysicians’Desk Reference 1754(58th Ed. 2004). at 42PatientInformation,Duragesic(FentanylTransdermal System)(May 2003),availableat htW/www.fda.~ov/cder/Dediatric/labels/Fentanvl.Ddf at 28. 43Kee GhanMoon dzHoward I. Maibach,Percutaneous Absorptionin DkeasedSkin: Relationshipto the Exogenous Demutoses, in EXOGENOUS DERMATOSES: ENVIRONMENTAL DERMATITIS17-226(Torkil Menne& 2 Howard I. Maibacheds., 1991). 44Petition at 7.


labeling -- warns that “[tlhere is a potential for temperature-dependent increasesin fentanyl releasedfrom the system.“5 In fact, both the “Warnings” and the “Precautions”sectionsof the labeling contain a detailed instruction to clinicians, in capitalizedtext, stating: “All patients and their caregiversshould be advisedto avoid exposingthe DuragesicQapplication site to direct external heat sources,such as heating pads or electric blankets,heat lamps, saunas, tubs, and hot heatedwater beds, etc., while wearing the sy~tem.“~ The labeling further advisesof the possibility of an increasein serum fentanyl concentrationsfor patientswith a fever and cautions clinicians to monitor any patientswith an elevatedbody temperature.47 Finally, the Duragesic@ patient information, which is provided directly to patientsto whom fentanyl transdermalsystems are prescribed,containsthe following warning: “Do not use heat sources such as heating pads, electric blankets, heat lamps, saunas, hot tubs, or heated waterbeds. Do not take long hot baths or sun bathe. All of thesecan make your temperaturerise and causetoo much of the medicine in DURAGESICB to passinto your body.‘ Again, the labeling and patient 48 information for Noven’ generic fentanyl transdermalproduct will contain identical warnings s and instructions. As this external factor (whether from an external sourceor from an increasein body heat) impacts the rate of releaseregardlessof the delivery system,there is no reasonto classi@matrix systemsas a different dosageform. Indeed,ALZA determinedthat the rate of delivery was not any greaterfor its own matrix systemwithout a rate-controlling membranewhen exposedto heat

*’SeePhysicians’Desk Reference 1753(58th Ed. 2004). at 46Id. 47Id. ‘ PatientInformation,Duragesic(Fez&my1 * Transdermal System)(May 2003),availableat htto:/lwww.f~~ov~~l~~~c~a~ls~en~vl,~f at 26 (emphasis original). in


than the rate for its DuragesicB reservoir system.4gThus, there is no evidenceto suggestthat exposureto heat rendersthe Noven matrix systema different dosageform from the DuragesicQ reservoir transdermalsystem. 3. The Clonidine Citizen Petition Does Not Presenta Valid Comparisonto the Safety ConsiderationsRelevantto FDA’ Review of Fentanyl s TransdermalMatrix Svstems.

ALZA’ referenceto the pending Citizen Petition (Docket No. 2001P-0470)regarding s transdermalclonidine is also unavailing. Drawing a parallel betweenthe generic clonidine transdermalmatrix systemand a fentanyl transdermalmatrix systemis inappropriatefor a number of reasons. Fentanyl and clonidine are very different drugs. Clonidine has a narrow therapeutic index,5o defined as a less than 2-fold difference in median lethal dose (LD50) and median whereasfentanyl has a therapeuticindex of at least 117.52 effective dose(ED50) values,51 Transdermalabsorption,like GI absorptionor any other physiologic property, demonstrates s variability that can be describedby a normative plot. Fentanyl’ therapeuticindex of at least 117 demonstrates that, even accountingfor the range of permeability of skin found in the general population, it is vastly safer than clonidine throughout that range of skin absorption,thus making ALZA’ comparisonof the two drugs misplaced.53A more appropriateanalogy can be made to s

49Petition at 8. ” SeeFDA, Guidance Industry: ImmediateRelease for Solid Oral DosageForms: Scale-Upand Post-Approval Changes:Chemistry,Manufacturingand Controls,In Vitro D&solution Testing,and In VivoBioequivalence Documentation (Nov. EM), available at h#p://www.fda.p~i~~/c~5.bdf(listing A-l Clonidine HydrochlorideTabletsand ClonidineTransdermal patchesas uarrow therapeutic raugedrugs). 5’21 C.F.R. 0 320.33(c). a The therapeutic indices for fentauylarebetween117 and 2OOas analgesic.SeeSummaryBasis of Approval. an Duragesic.October1990(obtainedvia FOIA request). 53Subsequent the submissionof the ALZA Petition to FDA, the law firm of London& Mead submittedits own to Citizen Petition relatedto FDA approvalof ANDAs for fentanyltransdd systems. Citizen Petitionby London


the drugs for which FDA has approvedNDAs in a matrix transdermalform. Theseproducts -nitroglycerin, nicotine and estradiol-- have broadertherapeuticindices and are thereforemore comparableto fentanyl than is clonidine. Moreover, unlike fentanyl, clonidine is well known to be irritative to the ~lcin.~~ Since compromisedskin is potentially more likely to have a higher rate of absorption-- leading to contraindicationin the labeling of transdermalproducts -- clonidine has an inherent characteristic that might lead to overdosage when delivered transdermally. This is not true for fentanyl. Thus, any concernsregarding safety of a clonidine patch have little, if any, bearing on the determinationof whether a fentanyl matrix systemmeetsthe criteria for ANDA approval.

ALZA has failed to advanceany valid scientific evidenceto support its speculationthat the matrix delivery systemutilized in Noven’ generic fentanyl product has a greaterpotential s for abuseand misusethan its brandedreservoir delivery system. Instead,the scientific data previously submittedby Noven in support of its ANDA, as well as the information and analysis presentedherein and in the CommentsNoven submittedin responseto the Brookoff Petition, establishthat Noven’ fentanyl transdermalmatrix delivery system(i) is not subjectto the risk of s abusehypothesizedby ALZA and (ii) is, in fact, substantiallyless likely to be subjectto potential abusethan is Duragesic@.The relief ALZA seeksin the form of an RMP for generic -- and only generic -- fentanyl transdermalsystemsis neither legally permissiblenor scientifically justifiable. Similarly, ALZA has failed to provide any valid scientific or legal basis for FDA to

& Mead,Docket No. 04P-054OKPl (submittedto PDA on or aboutDecember 7,2004). The London& Mead Petition dealsmore particularly with the issueof skin permeability,andNoven will be submittinga separate response that docketthat addresses issuein greaterdepth. in the
54See Physicians ’Desk Reference at 1002-03 (58th Ed. 2004).


determinethat the matrix patch should be classified as a different dosageform fi-om the reservoir patch. Accordingly, Noven respectfully requeststhat the Citizen Petition be denied and that FDA proceedto approval of Noven’ ANDA for its generic fentanyl transdermalmatrix patch so s that Noven may competewith the brandedproduct when ALZA’ monopoly endson January23, s 2005.




Respectfully submitted, NOVEN PHARMACEUTICALS, INC. 11960 SW 144* Street Miami, Florida 33 186 (305) 964-3132 A By-. -M$ Eduardo G. Abrao, M!D., PhD d Vice President Clinical Development & Chief Medical Officer

Chief Technical Officer

Director, Clinical Research

Of Counsel: Jeffrey F. Eisenberg, Esq. Vice President & General Counsel


Gary J. Buehler, Director, Office of Generic Drugs Gerald F. Masoudi, Acting Chief Counsel Elizabeth H. Dickinson, Associate Chief Counsel