WESTON LIBRARY
`
`JUN O 8 2000
`
`JS/120 CLINICAL SCIENCE CENTER
`
`
`
`
`600 HIGHLA'l} AVE l.'I\D1SON WI 53792
`
`Bausch Health Ireland Exhibit 2007, Page 1 of 17
`Mylan v. Bausch Health Ireland - IPR2022-01105
`
`

`

`T111·wp,·1111c-Dmi: Mot1i1ori11.�
`
`22:330 )..5 0 2000 Ltppincou William,
`· \ il�llh, Inc� Philadelphia
`
`Recommendation for Bioequivalence Te ting of Cyclo porine
`
`
`G neric R vi it d
`
`we Chri tian ,* M. Roy ir t, t and e lie Z. Benet,*
`
`'Departme111 n.f Bioplwr111ace11tical Sci,mce.r, School of Pharmacy, Uni"':r.rity n/ a/(fomia, ,111 Francisco, CA; tUniver.�if)• of
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Cinc.:il!lwri Medical Ce/lier, Ci11ci111w1i. OH
`
`Summary: The immuno uppres ·ant cyclo.
`pnrine i-; g nerally con ider a critical ­
`
`
`
`
`
`
`
`
`
`d o e drug. The validity o f t:tndard criteria t o e rnblish bioequivalcnce between cy­
`
`cd. Re mmendations included
`
`
`
`closporinc fo1TT1ulations ha recently been hallen0
`
`
`
`
`
`e. tabli hment of individual bioequivalenc rather than average biocquivalence, estab­
`
`
`
`li hment of biocquivalence in transplant patient. and in ubgroup · knO\ n to be poor
`
`nb orber , as well long-term efficac: ond safety studies
`
`in transplant p ticn .
`
`
`
`
`H wevcr. ::it th m oment individual bioequivalcnce ·s a theoretical concept, the prac­
`
`
`
`
`tical benefits of ,,hich ha,·e not tati tically ccn pr vcn. The proposed patient phar­
`
`
`to r quire an unreali tically high number of
`
`
`macodynamic s1udies can be expected
`
`
`
`
`
`
`subject lo achieve ufficient statistical power. It i well establi hed that the common
`
`
`
`ration-guidcd d ing of cyclosporine cf!iciemly compcns.ue.
`
`practice of bl od-concent
`
`
`for interindividanl nnd iniraindividuul variability and flows for. afoly witching
`
`cy­
`
`
`clo p rine formulation bioinequivalent as andimmune an eoral. R nl tu ies
`
`
`
`omparing the gencri cycl porin fonnulation ang ya with eoral. inclu ing in­
`
`dividunl bioequiv3lencc. bioequival nee in lran plant patients. and I ng-�rm safety
`
`
`t conclude
`
`
`after switching from and immune to ang •a, confirmed that it wn valid
`
`
`
`
`
`
`biocquivalencc of both cyclo·porine Ii nnulations ba.sed on standard average bio­
`
`
`
`
`
`
`equivalcncc criteria. Pre cnt FDA guidclin . for approving biocquiva.lcn c can be
`
`
`
`
`considered adequnle and uflici nt for generic cyclo porinc formulation.. c Word :
`
`
`
`yclo porine-Cyclo ·porine generics-Bioequivalence-lndividual bioequiva­
`lcnce-Therapeutic drug monitoring.
`
`Mostly o a result of the i troduction of the undeca­
`
`
`eiving immuno uppres i e drug regimen based on cy­
`
`
`
`peptide cyclosporinc as immun upprcssant, graft and
`
`
`closporine. Worldwide ale· of the innovator' cyclo-
`
`
`
`
`patient sur ival have ignificantly improved during the
`
`
`
`porine formulation Sandimmune and eornl ( ovarti
`
`last two decade and u·an plantation is an e tabli hcd
`
`
`at 1.3 Pharma, Ba. el, witzerland) were stimared
`
`
`
`standard procedure at mo t large medical center . How­
`
`billion in I 997. In the United tates, the innovator's
`
`ever there ar considerable co t for immuno uppre sive
`
`patent protection e ·pirc after 17-20 year and other
`
`
`
`therapy requiring life-long maintenance to pre ent the
`
`
`companie · are then free t manufacture interchange­
`
`
`
`transplant organ from being rejected (l,2). In the United
`
`
`able generi product . Novartis' comp ition of math.:r
`
`patent on cy lo porine expired in th nited tare in
`tates and Europe there ar more than 200,000 tran plant
`
`
`recipient requiring daily immuno uppre ive therapy
`
`
`September 1995. One generic cyclo potine formulation.
`
`SannCya ( angStm edical, San Mateo, CA. SA 1,
`for the re t of their live , the majority of whom are re-
`
`ha recently be n appro ed by the United tat Fo d
`
`and Drug Administration (FDA). Others have filed for
`Received June IO, 1999; accepted
`approval.
`December 30, 1999.
`Addrc�s 1:orrcspondcnce and reprint requeqs 10 Leslie Z. Benet,
`
`
`
`
`
`In 1984 th Drun Pri e Competition and Term Re. -
`
`
`
`Ph.D., Pmfc.. or, Department of Biopharm. cuti nl Sci n<:c�, School
`torati n ct (3) all -.: ed the FDA to use a implificd
`
`of Pharmacy. Univcr it)' of :ilifurnia. Snn ran i co, 33 Parna u
`
`
`ve. Room -68, San Franci ,:o, CA 94143-0446
`
`
`
`approval proces for generic drug products, the o-callcd
`
`330
`
`Bausch Health Ireland Exhibit 2007, Page 2 of 17
`Mylan v. Bausch Health Ireland - IPR2022-01105
`
`

`

`CYCLOSPORINE BIOEQUJVALENCE
`
`331
`
`abbreviated new drug application (ANDA) (4). The
`FDA' s approval pr0ce s of generic drugs evaluates
`diemistry, manufacturing and controls, in vivo bio­
`equivalence, labeling, in vitro dissolution if applicable,
`:md include inspection and auditing of all facilities (5).
`Because the efficacy and safety of an innovator s drug
`has already been e tablished, the FDA regulations are
`promulgated based on the belief that there i no reason to
`repeat the same studies with the generic version of the
`drug that contains exactly the same m lecular entity as
`the innovator's producL. Because of the lower cos , of
`development and competition in rhe market generic
`drugs usually ell for significantly less than the price of
`the innovator' product before the availability of gener­
`ics. It is generally agreed that the prescribing and use of
`generic drugs lead to considerably reduced cost. Generic
`drug also have the potential to improve the quality of
`care. Lower-cost alternatives may improve adherence to
`therapies for patients who cannot afford innovator drugs,
`and these alternatives provide an increased duration of
`therapy for patients with capped medical benefits. Dur­
`ing the last 27 years, the FDA has approved more than
`5,000 generic drugs for marketing in the United States
`�5 . To date, the FDA is not aware of any validated study
`of an FDA-designated equivalent generic product that
`met FDA pecifications but that was not equivalent to the
`conesponding innovator' product (6,7 . In addition, the
`FDA' s investigation of single cases of decreased efficacy
`or increased toxicity never revealed problems attributed
`to substitution of one approved product for another thera­
`peutically equivalent product (7). In spite of this excel­
`lent safety record, there is a great reluctance by many
`clinicians to use generic equivalents for so-called "criti­
`cal-dose drugs." Although there is no official definition
`for "critical-dose ' or "nanO\v-therapeutic-index" drags.
`and no general consensus as to which drugs fall within
`
`this category (8), bioequivalence-related issues of criti­
`cal-dose drugs have been discussed intensively. Benet
`and Goyan (9) defined narrow-therapeutic-index drugs
`as "those for which small changes in pharmacokinetic
`response lead to marked changes in pharmacodynamic
`response." Accordingly, cyclosporine is generally re­
`garded as a typical critical-dose drug (10-15). Bio­
`equivalence testing procedures, especially in the case of
`critical-dose drngs, have been criticized in the past for
`many reasons, most of which potentially apply to cyclo­
`sporin (9,10,12,13). A fundamental problem is the defi­
`nition of bioequivalence, which is based on the assump­
`tion that bioavailability (rate and extent) is a valid sur­
`rogate for efficacy and safety (16, 17). This requires a
`clinically significant association between blood/plasma
`concentrations and pharmacodynamic effects that is not
`necessarily always the case. However, for cyclosporine
`the relationship between pharmacokinetics and safety
`has been extensively studied and provides the basis for
`the generally accepted blood-level-guided dosing regi­
`mens. Several other potential issues regarding the inter­
`changeability of cyclosporine formulations are of con­
`cern to clinicians. There is doubt that the results of piv­
`otal bioequivalence studies that are conducted in healthy
`volunteers are extrapolatable to transplant patients who
`exhibit several factors affecting cyclosporine pharmaco­
`kinetics that are not present in healthy volunteers (see
`below and Fig. 1). This applies especially for subpopu­
`lations of patients who are known poor absorbers. Intra­
`individual variability of cyclosporine is a critical clinical
`issue that has been associated with acute and chronic
`rejection (18,19) and cannot be addressed by pivotal
`healthy volunteer trials. This translates into suspicion
`that standard bioequivalence testing may not be a valid
`approach to establishing long-term safety and efficacy in
`transplant patients.
`
`a,rc
`disea c
`race
`<liet
`
`time after
`clinical
`pre-e, istmg
`�tors
`transplantation
`statu
`�
`/ rejection
`-------::.
`infection
`liver function
`gut function
`
`activity of:
`•CY P3A enzymes
`•ABC protein
`transporters
`/
`concomitant ' transplant
`..._ ___________ ___.
`liver
`CYP3A/ ABC protein
`organ
`therapy
`•induc rs
`heart
`•inhibitors
`small bowel
`
`FIG. 1. Factors potentially affecting cyclo­
`spori ne phnrmarnkin.ctics in transplant
`pi1tien1S.
`
`Ther D111g Monil. \lu/, 22, No. 3, 2000
`
`Bausch Health Ireland Exhibit 2007, Page 3 of 17
`Mylan v. Bausch Health Ireland - IPR2022-01105
`
`

`

`332
`
`U. CHRISTIANS ET AL
`
`TABLE 1. Comparison of guidelines and recommendations to establish bioequivalence and to switch between
`cyclosporine formulntions
`Johnston et al., 1 99710
`Sabatini et al., 1999 13
`Demonstration of individual
`Validity of average
`bioequivalence should be
`bioequivalence questionable
`mandatory for FDA approval
`
`Recommendation
`
`Average/individual
`bioequi valence
`
`Kahan, 1 99914·:?.1
`Average bioequivalence is a
`valid approach to establish
`interchangeability, individual
`bioequivalence should be
`demonstrated for the fust CsA
`generic approved
`Recommended for first CsA
`generic approved
`Recommended for first CsA
`generic approved
`
`Should be required for all CsA
`generics
`Should be required for all CsA
`generics
`
`Should be required for FDA
`approval of all CsA generics
`Should be required for FDA
`approval of all CsA generics
`
`Should be required (study period
`>3 months)
`Not addressed
`
`Not addressed
`
`Yes
`
`6-months pre-marketing
`follow-up
`Not required
`
`Bioequivalence studies in
`paticnlS af1cr 1ransplantation
`Biocquivnlencc studies in
`subpopu1a·1ions that are poor
`absorbers ,
`Long-term efficacy and safety
`studies in transplant patients
`Physicians and patients must
`approve switch of CsA
`formulations even if
`bioequivalent
`
`CsA, cyclosporine.
`
`The question has been raised by several authors ( 10,
`12,1 4,20) as to what extent the standard bioequivalence
`criteria used by the FDA and most drug agencies in other
`countries address these concerns and the sufficiency of
`these criteria to establish the safety of substituting cy­
`closporine formulations. This has also been discussed in
`recent meetings ( 1 3,21 *). This has resulted in several
`different and sometimes contradictory guidelines and
`recommendations (Table 1). It was our goal to critically
`review cyclosporine bioequivalence issues and the dis­
`cussed recommendations in light of bioequivalence and
`clinical data that is presently available for several generic
`cyclosporine formulations and in light of the extensive
`experience with switching transplant patients between
`the innovator's bioequivalent cyclosporine formulations
`as well as between the bioinequivalent Sandimmune and
`Neoral formulations.
`
`CYCLOSPORINE FORMULATIONS
`
`Recognizing the limitations of the original cyclospor­
`ine formulation Sandimmune, a crude oil-in-water drop­
`let mixture (22), the innovator (Novartis Pharma, B asel,
`Switzerland) developed a microemulsion preconcentrate,
`Neoral, that improved emulsification and dispersion of
`cyclosporine in the small intestine and resulted in better
`and more reproducible absorption (23,25). From the be­
`ginning, Neoral was developed to increase cyclosporine
`bioavailability and, therefore, to be bioinequivalent (i.e.,
`
`*Generic Immunosuppressants: Should you be worried? Transplan­
`tation Society sponsored symposium. Montreal, Canada, July 1 2. Pre7
`sentations were published in Transplant Proc 1999; 3 1 [supplement].
`
`Ther Dmg Mo11it, Vol. 22, No. 3. 2000
`
`suprabioavailable) to Sandimmune ( 10,20,24). In fact,
`Sandimmune and Neoral should be considered different
`drug products (20).
`In healthy volunteer studies (25,26) as well as in clini­
`cal studies in transplant patients (23-25,27) and psoriasis
`patients (28,29), Neoral cyclosporine pharrnacokinetics
`differed from those of Sandimmune, yielding increased
`maximum blood concentration (Cmax), decreased time to
`reach Cmax (tmaJ, and increased area-under-the-time­
`concentration curve (AUC) (23). Depending on the dose,
`the relative bioavailability of Neoral in healthy volun­
`teers was 1 .7-fold to 2.4-fold and the Cmax 1 .9-fold to
`2.1-fold higher than after the same Sandimmune cyclo­
`sporine dose (26). In de nova recipients of kidney trans­
`plants, depending on the time after transplantation, dose­
`normalized AUCs were 32-63% higher than in Sandim­
`mune-treated patients (27). The mean increases of AUC
`and Cmax of 39% and 15%, respectively, in stable recipi­
`ents of kidney transplants after switching from Sandim­
`mune to Neoral (30) were smaller than in the healthy
`volunteer studies (26). Although based on healthy vol­
`unteer studies, a conversion factor of 0.6 (Neoral:Sand­
`immune) was estimated, tr�nsplant patients were
`switched 1: 1 (25). In a clinical study in 55 stable recipi­
`ents of kidney transplant, switching from Sandimmune to
`Neoral on a 1: 1 basis resulted in 22% higher cyclospor­
`ine trough blood concentrations (3 1). However, patients
`with higher cyclosporine doses before conversion from
`Sandimmune to Neoral are more likely to require dose
`reduction in the postconversion course. When switched
`from Sandimmune to Neoral, good absorbers remain
`good absorbers whereas poor absorbers become good
`absorbers (32). The higher bioavailability and different
`
`Bausch Health Ireland Exhibit 2007, Page 4 of 17
`Mylan v. Bausch Health Ireland - IPR2022-01105
`
`

`

`CYCLOSPORINE B!OEQUIVALENCE
`
`333
`
`and lower intraindividual pharmacokinetic variability,
`Neoral is generally considered to have proven benefits to
`patient care over Sandimmune (2, 10,24,27).
`In October 1998, the FDA approved SangCya (Sang­
`stat Medical, Menlo Park, CA, USA) as the first generic
`cyclosporine formulation in the United States. SangCya
`is a nano-dispersion formulation based upon Sangstat's
`CPLF formulation technology (37). Bioequivalence with
`Neoral was not only established in pivotal healthy vol­
`unteer studies (38), but also in recipients of kidney and
`liver transplants (39,40) (Table 2, Fig. 2). In addition,
`individual bioequivalence between SangCya and Neomi
`was demonstrated (4 1) (Table 3, see below) following
`the draft FDA procedures ( 1 1,42). Safety and efficacy of
`SangCya was established in patients with kidney grafts
`during a 9-month observation period (43).
`Healthy volunteer studies demonstrating bioequiva­
`lence with Neoral (Table 2) have been published for two
`other generic cyclosporine formulations, Neoplanta
`(Hanmi Pharmaceutical, Seoul, Korea) ( 44,45) and Ci­
`pol-NR (Chong Kun Dang, Seoul, Korea) (46). Like
`Neoral, both are microemulsion formulations (46,47).
`The difference between Neoplanta and Neoral is that
`Neoplanta uses dimethyl isosorbide instead of ethanol as
`the solvent (48). In de novo recipients of renal trans­
`plants, Neoplanta and Neoral (n = 20 for each group)
`showed similar efficacy in preventing graft rejection and
`similar tolerability (48).
`
`pharmacokinetic pattern of Neoral raised several safety
`concerns that required clarification in clinical studies
`t23-25). The high cyclosporine Cma1- after Neoral was of
`·pecinl concern because high cyclosporine C,mi., values
`have been related Lo short-term renal vasocon triction
`and possibly chronic cyt:losporine nephropathy (33 ). An­
`other concern was the higher total expo ure of patients
`during conversion from S andimmune to Neoral (23). Tbe
`conversion protocol recommends starting Neoral at the
`preconversion dose ( I : 1 conversion) with subsequent
`do e adju tment according to cyclosporine trough blood
`concentrations. It was nece sary to as um
`that the
`greater exposure to cy losporine from the microemul ·ion
`formulation might increase the nephrotoxic risk. In fact
`adverse event such as hyperten. ion, nephrmoxicity, and
`acute rejection have been reported after conversion (30).
`However, as of today, despite the two products' signifi­
`cant pharmacokinetic differences, clinical studies have
`established a safety and tolerability profile of Neoral
`omparahle <o thaL of Sandimmunc (24). Long-term stud­
`ies did not sho,. ally st:.aistically significant differences
`between recipient of kidney tran. plants treated with
`Sandimmune and those treated with Neoral in terms of
`safety, including creatini ne concentrations, patient and
`graft survival, as well as the incidence of acute rejection
`(23,24,27 ,34,35 ). This is not surprising: because of the
`drug' s highly intraindividually and interindividually
`variable pharmacokinetics and narrow therapeutic index,
`cycl�sporine doses must be adjusted according to cyclo­
`VARIABILITY OF CYCLOSPORINE
`sporine blood concentrations (36). Regular therapeutic
`PHARMACOKINETICS
`drug monitoring is required, and the cyclosporine con­
`The significantly lower pharmacokinetic variability of
`centrations are kept in a narrow target concentration
`range that is independent of the cyclosporine formula­
`cyclosporine after administration of Neoral compared to
`tion. However, because of its improved dose linearity
`Sandimmune is commonly regarded as the major im-
`TABLE 2. Comparison of the res11/ts <Jf bioequivalmce studies in healthy ro/1111teers and patients who have fwd a
`tra11splalllation with cydmpori11efor111ulations (lest) /Jioequiva/ent to Neorcrl (referencer
`AUC ratio (%)
`c"'"' ratio (%)
`
`Cyclosporine
`
`Test Formulation
`SangCya
`SangCyat
`SangCya
`SandCya
`SangCya
`Neoplanta
`Cipol-N
`SangCya
`SangCyu
`
`Subjects
`
`Fasted male healthy volunteers
`Fasted male and female healthy volunteers
`Fasted/fed male healthy l'olunteers
`Fasted female healthy volunteers
`Fasted male Afric.in-Amcrican volunteers
`Fasted male Korean healthy volunteers
`Fasted male Korean healthy volunteers
`Kidney transplant patients
`Liver transplant patients
`
`36
`20
`24
`28
`1 0
`2-+
`24
`32
`26
`
`Point
`Estimate
`
`99
`95
`97
`92
`96
`97
`l 03
`90
`86
`
`90% Cl
`
`97-IO-+
`9()-101
`9 1 -104
`87-IO0
`8 1 -108
`90- 1 0 1
`I 0(l- 106
`84-102
`8 1 -106
`
`Point
`Estimate
`
`90% CI
`
`Ref.
`
`99
`97
`1 00
`95
`90
`99
`1 00
`94
`95
`
`97-103
`92- 1 02
`96-- 105
`92- 102
`83-96
`94- 1 02
`96-104
`86-106
`89-109
`
`38
`41
`3 8
`38
`38
`45
`46
`39
`40
`
`s; The AUC ratio in he:1llhy volunteer swdici is bascn upon tht: /\UC0_'l_, the AUC ratio studies on the AUC0_, in patients :-iftcr transplant.
`Ncnpl ::i.nrn�l and
`ipol- -'", like Ncor:i.J�. u.rc microcmul sion cydospori11e formulations, whereas SangCya is a nano-dispersion formulation based
`upon s�n�s1:1t's PLr- lormulation 11:chnology37,
`,· Analy�b nf individual bioequivnlcnt:e see T�blc 3.
`Cl, confidenec interval.
`
`Ther D111s Mo11i1, \ lo/, :?2. No. 3, ::!000
`
`Bausch Health Ireland Exhibit 2007, Page 5 of 17
`Mylan v. Bausch Health Ireland - IPR2022-01105
`
`

`

`334
`
`� 220
`�
`0
`:.:; � 200
`Q)
`C
`� 1 80
`
`� 1 60
`2
`C:
`d 1 40
`::>
`<t:
`
`1 20
`
`1 00
`
`80
`
`60
`
`U. CHRISTIANS ET AL
`
`FlG. 2. Comparison of biocquivalcnce of dif­
`ferent cyclMporiae formulati ns in health}' vol­
`unteers and stable kidney transplant patients.
`The bars represent the 90% confidence intervals
`of the AUCo..., test/reference ratio and the lines
`across the bar represcnL the polnL estimuccs. The
`dolled line repre ents cnrnplete equivalence
`( 1 00%), whereas t.hc cl!IShed llnes nre at 1 25%
`and 80'¼-, the biocquivalence ncccptnnce limits.
`Data is taken from references 38,39.
`
`Test/ Reference,
`study subjects
`
`en
`
`en
`
`en
`en
`c c �j
`Cl)
`-�
`ro �
`::J -
`0.
`0.
`E g_
`�c c E -·- C
`i ru
`_:- cu
`-o ro
`..._ ro
`� o.. § o..
`o -Cl) C
`Cl) 0.
`o -
`o en CJ) �
`Cl) C
`-- ro
`z � -a; � >. >,
`-- ro
`z �
`n, ...
`Cll L..
`>, ......
`z �
`�;_
`� �
`t> ..c o, ......
`(.) >,
`---
`CJ) Cl)
`0) Cl)
`C ro Cll Cl)
`(1) -0
`cc -0
`cu -0
`c c o c c c
`(f) .r:.
`(/) �
`Z '.52.
`Cl) '.52.
`
`TABLE 3. Comparison of intrasubject variability* and individual bioequivalence of Sa11gCya (test) a11d Neoral (refere11ce/ 1
`Ratio
`1Jpper 95%
`confidence
`(95% confidence
`interval for 81 t
`interval)
`1 .277
`
`Parameter
`Cma, [µg · L-1]
`CV
`AUC0-�• h (1.1.g · L-1 . h)
`CV
`AUC0_ [µg · L-1 • h]
`CV
`
`SangCya
`
`0.0235
`15.4%
`0.01 1 1
`1 0.5'«
`0.0 1 27
`1 1 .3%
`
`Neoral
`
`0.0327
`1 8.2%
`0.0124
`1 l .2'ii
`{J.008 1
`9.0%
`
`0.72
`(0.22-1 .76)
`0.89
`(0.36-2.20)
`1 .56
`(0.36-3.83)
`
`p-va:lue
`0.50
`
`0.84
`
`0.43
`
`1 .009
`
`0.935
`
`cah:ulatcd f llowmg rhe procctluJ"c dl!scribcd by Liu. lntra�ubjcct v:irinhiliry bet, ccn
`lnu ubjecl 1·ariabili1y \\
`:111g )'11 te. l) nnd cor.il
`crere.rcncc wu. C(lmp:11,.,'t.l u�ing the likellho
`ra1tu ;x1 1 1. lnu subjc�1 vari.tbifoy i� h. � d uu tile logarithmit: btalc
`t Bioe.qui vulcm c wns 1cccp1eo when the upper 95<"' cunfidcncc imcrv:11 wn :s lhc indivitlunl bincquivnlcncc li111it 01• which w 1� col ulntcd :II 2.1-i.5
`using I I < t�lrnp mcthotl (2000 :nrnplcsJ.
`C . coef 1cicn1 f imrasuhjcct 1111fability: C,,,,,,, maxim11111
`
`lo<Ki coni:cntr. ri1 n,
`
`
`
`TJ,a Dmg Mo11it, Vol. 22, No. 3, 2000
`
`Bausch Health Ireland Exhibit 2007, Page 6 of 17
`Mylan v. Bausch Health Ireland - IPR2022-01105
`
`

`

`CYCLOSPORlNE B!OEQUIVALENCE
`
`335
`
`provement of Neoral over Sandimrnune ( 1 0, 13,14). Fluc­
`tuating cyclosporine blood concentrations have been as­
`sociated with chronic and acute rejection ( 1 8, 19 ,49). 1n
`comparison to Sandjmmune, the more consistent absorp­
`tion from the Neoral formulation may result in a reduced
`incidence of chronic rejection ( 1 8) and toxicity, it is ex­
`pected to make clinical management easier and safer
`(50), and it will reduce costs after transplantation (2).
`Demonstration of equivalent pharmacokinetic variability
`of generic cyclosporine formulations and Neoral has
`been a major concern ( 1 0,13,14,2 1).
`Factors· that play a major role in the low and variable
`oral bioavailability of cyclosporine include solubility,
`emulsification, countertransport of the drug by Pl 70-
`glycoprotein and other ATP-binding cassette (ABC) pro­
`tein transporters from the gut mucosa back into the gut
`lumen, and first-pass metabolism in the small intestine
`and liver.
`After administration of cyclosporine as the original
`Sandimmune formulation, absorption of cyclosporine re­
`quires the following subsequent steps: formation of an
`oil-in-water droplet mixture with gastrointestinal fluids,
`emulsification of this mixture by bile salts, digestion of
`the oil droplet, and solubilization of cyclosporine in
`monoglycerides and bile salts resulting in a mixed mi­
`cellar phase from which cyclosporine is absorbed
`(22,25). Emulsification by bile salts has been identified
`as the step that causes most of the variability in intestinal
`absorption of cyclosporine after Sandimrnune adminis­
`tration. This step is dependent on food intake, bile flow,
`and gastrointestinal motility (5 1). Microemulsion and
`nano-dispersion cyclosporine formulations are hypoth­
`esized to shortcut the critical emulsification step. In the
`Neoral microemulsion, cyclosporine is dissolved in a
`mixture of com oil mono-, di- and triglycerides, the hy­
`drophilic solvent propylene glycol, the surfactant poly­
`oxyl-40 hydrogenated castor oil, and the antioxidant DL­
`tocopherol (22). Upon contact with gastrointestinal fluid,
`a monophasic microemulsion is formed that has proper­
`ties similar to the putative mixed micellar phase from
`which cyclosporine is absorbed.
`Cyclosporine is a substrate of cytochrome P450 3A
`enzymes and the ATP-binding cassette transporter P170-
`glycoprotein (52-55). It is metabolized by CYP3A en­
`zymes in the small intestine to its major metabolites (56).
`In patients, metabolites were found to account for as
`much as 50% of the measurable cyclosporine derivatives
`in portal vein blood after cyclosporine instillation into
`the small intestine (57). In microsomes isolated from the
`duodenum of patients, cyclosporine metabolism varied
`IO-fold (56,58). A clinical study using intubation tech­
`niques to deliver cyclosporine to different parts of the
`
`gastrointestinal tract established a significant inverse
`correlation between cyclosporine absorption and Pl 70-
`glycoprotein messenger RNA at the administration site
`(59), suggesting that P l 70-glycoprotein-mediated intes­
`tinal countertransport significantly contributes to the in­
`complete absorption of cyclosporine. In a recent clinical
`study in stable recipients of kidney grafts (58), it was
`found that hepatic metabolism was responsible for 56%
`of the interpatient variability in apparent oral cyclospor­
`ine clearance and 32% of the variability in Cmax· After
`the liver effect was taken into account, the only other
`parameter significantly contributing to cyclosporine
`pharmacokinetic variability was intestinal Pl 70-glyco­
`protein, which was estimated to explain 17% of the vari­
`ability in apparent oral clearance and 30% of the vari­
`ability in Cmax (58). In the same study, cytochrome P450
`3A enzyme activities in the liver varied 3-fold and P l 70-
`glycoprotein in the small intestine IO-fold among pa­
`tients. These studies demonstrate that cytochrome
`P4503A-dependent intestinal and hepatic first-pass me­
`tabolism as well as P l 70-glycoprotein-mediated intesti­
`nal countertransport reduce the oral bioavailability of
`cyclosporine whereas hepatic metabolism and intestinal
`countertransport also contribute to its pharmacokinetic
`variability.
`
`AVERAGE BIOEQUIV ALENCE TESTING
`In the 1 970s it was recognized that, even when two
`drug products contained the same active component at
`the same dose, small changes in the product formulation
`could result in significant differences in oral bioavail­
`ability. Several cases of lack of effect or intoxication
`after administration of pharmaceutically equivalent ge­
`neric drug products were reported (60). Pharmaceutical
`equivalents contain the same active ingredient, are ad­
`ministered by the same route in the same dosage form,
`and are of identical strength and concentration (6 1).
`These experiences triggered an international effort to de­
`velop clinical and statistical procedures to establish bio­
`equivalence between pharmaceutical equivalents. Today,
`drug regulatory authorities in the United State.s (62), the
`European Community ( 17), and most other countries re­
`quire demonstration of average bioequivalence between
`the marketed and a generic drug product as the basis of
`approval. The rules to establish bioequivalence are basi­
`cally similar in most countries with only minor differ­
`ences. Bioequivalence studies typically aim to demon­
`strate that two pharmaceutical equivalents have similar
`pharmacokinetics (63). The standard bioequivalence trial
`is conducted according to a randomized 2-period cross­
`over design and includes from 12-36 healthy normal
`male adults with an appropriate wash-out between study
`
`Ther Drug Monit, Vol. 22, No. 3, 2000
`
`Bausch Health Ireland Exhibit 2007, Page 7 of 17
`Mylan v. Bausch Health Ireland - IPR2022-01105
`
`

`

`336
`
`U. CHRISTIANS ET AL
`
`periods. The key issue in bioequivalence testing is to
`demonstrate similar oral bioavailability. Because the
`pharmaceutical equivalents are orally administered, ab­
`solute bioavailability cannot be directly determined.
`Area-under-the curve (AUC) measurements serve as a
`surrogate for the extent of absorption; the maximum
`plasma concentration (Cma,) and the time of its occur­
`rence (tmax) together characterize the rate of absorption
`(64). Pharmacokinetic parameters used to establish bio­
`equivalence in the FDA and European Committee for
`Proprietary Medicinal Products (CPMP) guidelines are
`shown in Table 4. Test and reference product are con­
`sidered equivalent when the 90% confidence interval for
`the true formulation means (µ,10_/µ,rcfcrcnce) falls within
`the acceptance limits of 0.8-1 .25 (17,62). In practice, the
`confidence interval approach is carried out using log­
`transformed data (65). The 0.8-1.25 bioequivalence ac­
`ceptance range translates into a difference in rate and
`extent of absorption between the two drug products of
`-20% to +25%. These acceptance limits are based on the
`medical decision that a -20%/+25% difference in the
`concentration of the active ingredient in blood will not be
`TABLE 4. Pharm,r rcnki11eric parameters i11 the United
`Sw1e.� 11ml 'Ew·,,pcan g11irlcJi11es for hioequivalence testing 9s
`Recommended
`pharmacok.i netic
`parameters
`Single dose
`
`United States
`and Canada"'
`Cmux
`AUC0_,
`AUC0.�
`1112
`
`tnrnx
`
`Eu10pct
`cnrnx
`tma:-:
`AUC0.,
`AUC0_�
`t,/2:j: MRT:t
`Ae
`Ae0.�
`dAe/dt
`Css,,,ax
`Cssm in
`AUCT
`
`Multiple dose
`
`cmax
`Cm;n
`AUCT
`AUC0.�
`Cu,· DF
`
`lm.::ix
`
`* Food and Drug Administration
`t Committee for Proprietary Medicinal Products
`:j: Mentioned in the CPMP guideline 1 7 as optional parameters.
`Ae, cumulative urinary excretion from administration to the time
`point of the last measured concentration; Ae0_�, cumulative urinary
`excretion extrapolated to infinity; dAe/dt, urinary excretion rate;
`AUC0_,, area under the concentration time curve from administration to
`the time of the last measured concentration; AUC0.�, AUC extrapo­
`lated to infinity; AUCT, AUC during a dosing interval; Cmax• maximal
`blood/plasma concentration; Cssmax, maximum blood/plasma concen­
`tration at steady state; Cm;n, minimum blood/plasma concentration; C0 .. ,
`average blood/plasma concentration; Cssm,n• minimum blood/plasma
`concentration at steady state; DF, degree of fluctuation; MRT, mean
`residence time; 1112
`, blood/plasma concentration half-life; tmw time
`from administration to Cmnx·
`
`Titer Drug Monit, Vol. 22, No. 3, 1000
`
`clinically significant (6 1). It is important to recognize
`that it is the upper and lower limit of the 90% confidence
`interval for the trUe mean ratios and not only the mean
`ratio (point estimate) that must be within the bioequiva­
`lence acceptance limits (61). The 90%-confidence inter­
`val is a measure of total variability, which is influenced
`by both interindividual and intraindividual variability
`(1 1,66). Variability is a factor that has a significant im­
`pact on acceptance or rejection in average bioequiva­
`lence testing.
`It has been suggested that the standard procedures to
`establish bioequivalence may not be adequate for all
`drugs and that modified procedures and additional data
`may be necessary (9,60,63,67). Drugs for which the va­
`lidity of the standard approach for establishing bio­
`equivalence must be assessed and if necessary modified,
`are (1) those with a narrow therapeutic index, (2) those
`with high interindividual and intraindividual pharmaco­
`kinetic variability, (3) those for which pharmacokinetics
`does not correlate with pharmacodynamic effects, and
`(4) those with nonlinear pharmacokinetics and/or con­
`trolled modified-release formulations (60). The validity
`of standard average bioequivalence procedures to estab­
`lish bioequivalence of cyclosporine generics has been
`challenged (I 0, 1 3 ), mostly because cyclosporine has
`been classified as a narrow-therapeutic-index, highly
`variable drug ( 1 1 -14). A drug is commonly regarded as
`highly variable when it exhibits an intrasubject coeffi­
`cient of variance 2::30% as estimated by analysis of vari­
`ance (66,67). This criterion was clearly met by cyclo­
`sporine pharmacokinetics after oral administration of the
`original Sandimmune formulation. However, intrasu­
`bject variability of cyclospo1ine pharmacokinetics is for­
`mulation-dependent. Estimates of intrasubject variability
`in cyclosporine AUCs of 8% (68), 7% (69), 20% (70),
`and 9-21 % (7 1) after Neoral administration have been
`reported in patients with kidney transplants.
`The validity of average bioequivalence and the 0.8-
`1 . 25 acceptance range for narrow-therapeutic-index
`drugs has repeatedly been questioned. Tighter accep­
`tance criteria, such as an acceptance range of 0.9- l . l or
`the use of 95%- instead of the 90%-confidence intervals,
`have been proposed for narrow-therapeutic-index ·drugs
`(9) and are required by some drug regulatory agencies
`such as Canada's (72). In the United States it is believed
`that the present requirements to prove bioequivalence are
`already rigorous enough to prev