Chronic renal failure and disposition of
`hepatically metabolized drugs
`
`Ethnic differences in response to
`
`Fluvastatin influence
`on umbilical vein endothelium
`
`out'na.
`ona..
`ph <:n'macology
`t h er' ape uti. cs
`BML Flo Ot' 2 .
`UC San Diego
`Received on: 05-19-2000
`
`Boehringer Ex. 2020
`Mylan v. Boehringer Ingelheim
`IPR2016-01565
`Page 1
`
`

`
`This journal is included in the abstracting and
`indexing coverage of the Bio-Sciences Infor(cid:173)
`mation Service of Biological Abstracts, and is
`listed in Current Contents (Life Sciences), IN(cid:173)
`DEX MEDICUS/MEDLINE,
`International
`Pharmaceutical
`Abstracts,
`EMBASE/
`Excerpta Medica, Elsevier BIOBASE/Cur(cid:173)
`rent Awareness in Biological Sciences.
`All articles published in International Jour(cid:173)
`nal of Clinical Pharmacology and Thera(cid:173)
`peutics must be contributed to it exclusively.
`The following types of papers are acceptable:
`original papers, outline summaries of special
`problems, letters to the editor, short news
`items on new developments and other topics
`of current interest. All experimental data must
`be obtained from humans. Animal and in vitro
`investigations can be published only occa(cid:173)
`sionally and only under the condition that
`they are in close relationship to human phar(cid:173)
`macological investigations.
`The editor responsible reserves the right of
`selection from submitted manuscripts (after
`consulting at least two reviewers) and the
`right to make stylistic changes or abridge(cid:173)
`ments. The manuscripts may not be submit(cid:173)
`ted elsewhere for printing or publication; fol(cid:173)
`lowing acceptance the publisher acquires all
`copyrights. Rejected manuscripts are re(cid:173)
`turned to the authors.
`
`Publisher
`Dustri-Verlag Dr. Karl Feistle
`Bahnhofstr. 9, D-82041 Deisenhofen/
`Munchen, Germany, Tel. +49-89-613861-0,
`Fax +49-89-6135412
`email: customer.service @dustri.de
`
`Managing Editor
`Dr. Joyce Mayer, Dustri-Verlag, Dr. Karl
`Feistle, Bahnhofstr 9, D-82041 Deisenhofen(cid:173)
`Munchen, Germany, Tel. +49-89-613861-25
`email: dustri-jm @t-online.de
`
`Editor-in-Chief
`Prof. Dr. G. Hitzenberger, Kinderspital(cid:173)
`gasse 10/16, A-1090 Wien , Austria
`
`Deputy Editor
`PD Dr. B.G. Woodcock, Klinikum der JWG(cid:173)
`Universitiit, lnstitut fUr Klinische Pharma(cid:173)
`kologie, Theodor-Stern-Kai 7, D-60590
`Frankfurt/M, Germany, email: Woodcock @
`em.uni-frankfurt.de
`
`Editors
`Prof. Dr. D. Birkett, Clinical Pharmacology,
`Flinders Univ School of Medicine, Flinders Med
`Centre, Bedford Park, South Australia, 5042
`Prof. Dr. H. Derendorf, University of Florida,
`College of Pharmacy, Dept of Pharmaceu(cid:173)
`tics, Health Science Center, P.O.Box
`100494, USA- Gainesville, FL 32610-0494
`Prof. Dr. U. Fuhr, lnstitut fUr Pharmakologie
`der Universitiit zu Koln , Gleuelerstr. 24,
`D-50931 Koln
`Prof. Dr. M. Levy, Hadassah Univ. Hosp./
`Med., P.O.Box 12 000, IL-91120 Jerusalem
`Prof. Dr. N. Rietbrock, Am Sellsiekbach 52 ,
`D- 32657 Lemgo
`
`Prof. Dr. J.-P . Tillement, Univ. de Paris XII,
`Fac de Med de Creteil , 8, rue du General
`Sarrail, F-94010 Creteil
`Prof. Dr. T. K. Shibuya, Tokyo Medical College,
`6-1-1, Shinjuku, J- Shinjuku-Ku, Tokyo 160
`
`Associate Editor
`K.S . Estes, PhD , Clinical Pharmacology,
`8708 SW 42nd Place, Gainesville , FL 32608,
`USA, email : Kerry @global.net
`
`International Journal of
`
`clinical
`pharmacology
`and therapeutics
`
`© 2000 Dustri-Verlag Dr. Karl Feistle KG,
`Munchen-Deisenhofen, Germany
`ISSN 0946-1965
`All rights reserved (including those of transla(cid:173)
`tion into foreign languages). No part of this
`journal may be reproduced in any form - by
`photoprint, microfilm, or any other means nor
`transmitted or translated into a machine lan(cid:173)
`guage without the permission in writing of the
`publisher.
`Authorization to photocopy items for internal
`or personal use, or the internal or personal
`use of specific clients , is granted by Dustri(cid:173)
`Verlag Dr. Karl Feistle, provided that the base
`fee of$ 3.00 per copy is paid directly to Copy(cid:173)
`right Clearance Center, 27 Congress Street,
`Salem, MA 01970. For those organizations
`that have been granted a photocopy license
`by CCC, a separate system of payment has
`been arranged . The fee code for users of the
`is
`Transactional
`Reporting
`Service
`(0946-1965 $ 3.00).
`Registered names, trademarks , etc. used in
`this journal , even without specifi~ indication
`thereof, are not to be considered unprotected
`by law.
`Regarding a recommended dosage or an ap(cid:173)
`plication , authors, editors and publisher took
`highest possible care to ensure safety. How(cid:173)
`ever, the reader must consult the prescribing
`information of the medication to be used in or(cid:173)
`der to suit the needs of his individual patient.
`
`Volume 38 (12 issues per volume) will be
`published in 2000.
`Subscription price for 2000: OM 375.-,
`US $ 190.- I US $ 230.-lnstitution plus post(cid:173)
`age . Single copies: DM 40 .-, US$ 25.- plus
`postage.
`Available from all booksellers or directly from
`the publisher.
`
`Periodical postage paid at Jamaica, NY,
`U.S.A. Postmaster; Send address changes to
`International Journal of Clinical Pharma·
`cology and Therapeutics , Publications
`Inc., 200 Meacham Avenue,
`Expediting
`Elmont, NY 11003. Air freight and mailing in
`the U.S. by Publications Expediting, Inc., 200
`Meacham Ave ., Elmont, NY 11003, USA.
`Subscriptions are automatically canceled at
`the end of the year and recommence only with
`renewal. Prepayment is required on all orders.
`
`Printed in Germany by A. Butz, Haager
`Strar..e 8, D-81671 Munchen, Germany
`
`(20)
`
`Boehringer Ex. 2020
`Mylan v. Boehringer Ingelheim
`IPR2016-01565
`Page 2
`
`

`
`International Journal of Clinical Pharmacology and Therapeutics, Vol. 38- No. 5/2000 (245-253)
`
`Effect of chronic renal failure on the disposition
`of highly hepatically metabolized drugs
`
`R. Yuan1 and J. Venitz2
`
`1 CDER-OPS-OCPB-DPE-1, Food and Drug Administration, Rockville, MD, and
`2Department of Pharmaceutics, Virginia Commonwealth University, Richmond,
`VA, USA
`
`Abstract. Objective: The objective of this
`study was to investigate the effect of renal im(cid:173)
`pairment on the disposition of an extensively
`metabolized drug, i.e. , drug X. Dmg X has a
`hepatic extraction ratio of less than 0.1 and
`free fraction in plasma of less than l% in
`healthy volunteers. Methods: Pharmaco(cid:173)
`kinetic (PK) parameters of dmg X were ob(cid:173)
`tained from subjects with normal renal func(cid:173)
`tion (I, n = 6), as well as in subjects with mild
`(Il, n = 5), moderate (III, n = 7) and severe re(cid:173)
`nal impairment (IV, n = 5). Disease-PK mod(cid:173)
`els were developed to describe the changes of
`PK parameters with respect to renal function
`measured by creatinine clearance. While ex(cid:173)
`perimentally observed data are presented for
`dmg X, additional simulations were per(cid:173)
`formed for other drugs that are extensively
`metabolized (extensive metabolism is de(cid:173)
`fined as metabolism that accounts for more
`than 90% of total drug elimination). The sim(cid:173)
`ulated scenarios included drugs that have a
`low extraction ratio (ER) and with high
`plasma protein binding (PPB), low ER and
`with low PPB, highER and with high PPB, or
`high ER and with low PPB. Results: Systemic
`clearance of dmg X, a low ER and high PPB
`drug, in renal patients depended on the simul(cid:173)
`taneous effects of renal disease on protein
`binding and intrinsic metabolic clearance.
`Protein binding of drug X was related to
`creatinine clearance in an inverse hyperbolic
`relationship, while the unbound intrinsic met(cid:173)
`abolic clearance declined
`linearly with
`creatinine clearance. Because the disease ef(cid:173)
`fects on these two factors offset each other in
`terms of total systemic clearance, the lowest
`total systemic clearance was not observed in
`the severely renal impairment patients, but
`rather in the moderately impaired group. Ad(cid:173)
`ditional simulations showed that for low ER
`dmgs that are highly metabolized, the pattern
`and magnitude of systemic clearance change
`in renal patients depended on how the disease
`affected PPB and/or intrinsic metabolic clear(cid:173)
`ance. But the systemic clearance of high ER
`drugs would not be as susceptible to the effect
`
`of renal disease as that of low ER dmg. Con(cid:173)
`clusions : Chronic renal disease should not be
`considered as an isolated event that affects
`only renally excreted drugs. Uremia may also
`modify the disposition of a highly metabo(cid:173)
`lized dmg by changes in plasma protein bind(cid:173)
`ing and/or hepatic metabolism.
`
`Introduction
`
`Liver function as it relates to dmg metabo(cid:173)
`lism has generally been assumed to be un(cid:173)
`changed in patients with chronic renal failure
`(CRF) as compared to patients with normal
`renal function (NRF). Based on this assump(cid:173)
`tion, the disposition of highly metabolized
`drugs in CRF patients is expected to be simi(cid:173)
`lar to that in NRF subjects. This implies that it
`may not be important to prospectively study
`new drugs in CRF patients, if they are almost
`exclusively hepatically metabolized. How(cid:173)
`ever, mounting evidence has shown that CRF
`can change the non-renal clearance (i.e.,
`mainly hepatic metabolism) and thereby
`cause alterations in the disposition of highly
`metabolized dmgs as well [Lam et al. 1997,
`Touchette and Slaughter 1991 ].
`CRF is known to cause alterations in
`plasma protein binding [Matzke and Keane
`1989]. In plasma, the unbound fraction (fu) of
`a drug can either increase or decrease in CRF
`patients depending on the physico-chemical
`properties and binding characteristics of the
`drug. Acidic drugs that preferentially bind to
`albumin usually have an increased fu as a re(cid:173)
`sult of a qualitative change in the binding
`site(s), decreased semm albumin levels due to
`renal albumin loss, and/or an endogenous
`binding displacers that accumulate in uremia.
`The fu of basic dmgs, on the other hand, may
`
`Keywords
`renal - impairment -
`metabolism - binding
`
`Received
`August 20, 1999:
`accepted in revised form
`November 22, 1999
`
`Correspondence to
`Dr. R. Yuan
`Building 1-3C39,
`340 Kingsland Street,
`Nutley, NJ 07110, USA
`
`Boehringer Ex. 2020
`Mylan v. Boehringer Ingelheim
`IPR2016-01565
`Page 3
`
`

`
`Yuan and Venitz
`
`decrease because of increased serum a 1-
`acid-glycoprotein levels as the response to a
`chronic inflammatory disease state [Gibson
`1986).
`CRF may also result in decreased hepatic
`metabolic activity for a drug [Gibson 1986,
`Touchette and Slaughter 1991). Product inhi(cid:173)
`bition and reverse hydrolysis as a result of
`acyl-glucuronide migration are the proposed
`mechanism to explain decreased hepatic me(cid:173)
`tabolism in CRF patients for some drugs
`[Debord et al. 1994, Fillastre 1994], but they
`do not offer an explanation for the entire ob(cid:173)
`served reduction in hepatic metabolism in
`CRF patients. In an ex vivo single-pass perfu(cid:173)
`sion study in rat liver, Terao and Shen [ 1985)
`demonstrated the existence of endogenous
`molecule(s) circulating in the uremic blood
`that inhibited the intrinsic hepatic metabolic
`clearance of propranolol. In their study, the
`normal liver perfused with uremic blood had
`a lower metabolic activity than the normal
`liver perfused with nonnal blood, while the
`uremic liver perfused with normal blood had
`the same metabolic activity as normal liver
`with normal blood. Although the exact nature
`of the circulating metabolic inhibitor(s) h!J.s
`not · been identified, this hypothesis is often
`cited as the mechanism for the reduced meta(cid:173)
`bolic clearance in CRF patients [Lam et al.
`1997, Matzke and Keane 1989, Touchette and
`Slaughter 1991).
`For a highly metabolized drug that has
`low hepatic extraction ratio (ER), systemic
`clearance (CL101) of the drug is determined
`primarily by fu and intrinsic hepatic meta(cid:173)
`bolic clearance (CLint) [Wilkinson and Shand
`197 5). In this paper, we report a case where
`the systemic clearance of a drug (drug X) in
`CRF patients was affected simultaneously by
`CRF effects on both CLint and fu, leading to an
`unusual U-shaped relationship between CL101
`and creatinine clearance (CLcrea)- Under such
`circumstances, severely renally impaired pa(cid:173)
`tients and NRF patients on the two extremes
`had the same CL101 while only moderately im(cid:173)
`paired patients had a reduction in CL101 . We
`developed a disease-PK model and conducted
`a series of simulations to illustrate different
`scenarios on how CL101 of highly metabolized
`drugs may be affected by CRF.
`Drug X is a synthetic compound with a
`small molecular weight (MW- 300 Dalton)
`that is administered therapeutically as a short
`
`246
`
`i.v. infusion. Of the total administered dose,
`less than 1% is excreted as unchanged drug in
`urine (fe) and Jess than 10% in feces. Hepatic
`glucuronyl conjugation of the parent drug ac(cid:173)
`counts for less than I 0% of the dose excreted
`in urine, whileas the rest are oxidative metab(cid:173)
`olites formed by different cytochrome (CYP)
`pathways. The main circulating metabolite is
`formed by CYP2D6 hydroxylation. Drug X is
`a low hepatic extraction ratio drug with a total
`systemic clearance of 8 1/h. In healthy sub(cid:173)
`jects, 99.6% of the parent drug in plasma is
`protein-bound, mainly to albumin. Linear
`pharmacokinetics are observed for the parent
`drug over one half to two times of the pro(cid:173)
`posed therapeutic dose range.
`
`Methods
`
`Data resources
`
`The clinical PK study was part of a new
`drug application submitted to US Food and
`Drug Administration (FDA) for review. ln
`this particular clinical study (see below), the
`sponsor
`provided
`the
`estimates
`of
`pharrnacokinetic parameters for each individ(cid:173)
`ual subject such as fu, AUCo.inf• CL101 and
`CLcrea• but did not perfonn the analysis re(cid:173)
`ported here. The only pharmacokinetic esti(cid:173)
`mates extracted from the original submission
`were the total AUCo-inf• the unbound fraction
`in plasma fu, and the amount of parent com(cid:173)
`pound excreted in urine (Ae). These parame(cid:173)
`ters were taken from alJ subjects who com(cid:173)
`pleted the study without any violation of the
`protocol.
`
`Case
`
`The clinical study included subjects with
`various degree of renal function stratified by
`severity of CRF, based on the observed
`CLcrea· Data included . in our analysis were
`from 6 subjects with NRF (group l, CLcrea >
`80 ml/min), 5 with mild (group ll, CLcrea = 50
`- 79 mllmin), 6 with moderate (group Ill,
`CLcrea = 20 - 49 ml/min) and 5 with severe
`chronic renal impainnent (group lV, CLcrea <
`19 ml/min). AIJ subjects received a short in(cid:173)
`travenous infusion of the drug. Creatinine
`clearance for each individual was detem1ined
`
`Boehringer Ex. 2020
`Mylan v. Boehringer Ingelheim
`IPR2016-01565
`Page 4
`
`

`
`PK of highly metabolized drugs in renally impaired patients
`
`247
`
`from two 24-hour urine collections, and the
`plasma protein binding ofthe drug was exam(cid:173)
`ined by equilibrium dialysis of the pre-dose
`plasma samples fortified with radiolabeled
`drug. The plasma and urine concentrations of
`the drug were determined using a validated
`HPLC method with UV detection. The detec- ,
`tion limit for the parent drug was 2 ng/ml in
`plasma and in urine.
`
`Pharmacokinetic analysis
`
`Total plasma clearance of the drug (CLtot)
`was estimated by non-compartmental analy(cid:173)
`sis:
`
`CLtot = dose/AUCo.inf
`
`(Eq. 1).
`
`The unbound plasma clearance of the
`drug was derived from normalizing CLtot by
`fu:
`
`(Eq 2).
`
`The fraction of the drug excreted in the
`urine (fe) was obtained as fe = Ae/dose. Ae is
`the amount of drug excret~d in urine. Hepatic
`clearance of total drug (CLheptot) was calcu(cid:173)
`lated as:
`
`(Eq 3).
`
`Using the well-stirred hepatic clearance
`model developed by Wilkinson and Shand
`[ 1975],
`the unbound
`intrinsic metabolic
`clearance (CLint) was calculated as:
`
`CLint= CLhep X Qhcpi'(Qhcp X fu- CLhep x fu)
`(Eq 4)
`
`where Qhep is the hepatic plasma flow and
`is assumed to be unaffected by CRF [Leblanc
`et al. 1996]; fu is the unbound fraction in the
`plasma.
`After reviewing the data obtained for drug
`X, the following empirical disease-PK model
`was used to relate fu to creatinine clearance
`(CLcrea):
`
`fu = fumax- (fu max- fu min) X CLcre.f(CLcrea +
`RF so)
`(Eq 5).
`
`This relationship postulates an inverse hy(cid:173)
`perbolic relationship between fu and CLcrw
`where fu max was the maximum unbound frac(cid:173)
`tion of the drug in the plasma at the lowest re(cid:173)
`nal function, fu min was the minimum unbound
`fraction of the dmg in the plasma achieved in
`
`NRF, and RF 50 was the creatinine clearance at
`which fu was decreased to 50% of (fu max -
`fumin, see also Figure ld).
`The
`following empirical disease-PK
`model was used to relate CLint to CLcrea:
`
`CLint= CLint0 + Sint X CLcrea
`(Eq 6)
`where CLinto (intercept) is the intrinsic
`clegtrance in severe CRF patients who have
`virtually no residual renal function, and Sint
`(the slope) is the increment in CLint with in(cid:173)
`crease in CLcrea·
`Because of the common practice of using
`the values from healthy volunteers (NRF) as
`baseline condition when PK of a dmg in renal
`patients are evaluated, the model can be
`reparameterized to:
`
`CLint= CLintN- S'int X% of ~CLcrea (Eq 7)
`
`which assumes a decline in CLint that is
`proportional to the decline in CLcrea· CLintN is
`the intrinsic clearance in NRF and S'int (the
`slope) is the decrement in CLint with each per(cid:173)
`cent decrease in CLcrea· Note that the most se(cid:173)
`vere CRF patients, who have 100% decrease
`of creatinine clearance, would have a CLint=
`CLintN - S' int X 100.
`For dmg X, after calculations of the perti(cid:173)
`nent variables from the clinical PK data, i.e.,
`CLint (CLcrea) and fu (CLcrea), the following
`parameters were estimated in order to de(cid:173)
`on
`the
`effect of CRF
`scribe
`the
`fumax,
`fu min,
`pharmacokinetic parameters:
`Rfso, CLint0
`, Sint' as well as CLintN and S' int·
`
`Pharmacokinetic simulations
`
`With the parameters obtained from mod(cid:173)
`eling fu and CLint of dmg X (see above), the
`following three cases were simulated for
`dmgs with similar pharmacokinetic charac(cid:173)
`teristics as dmg X:
`renal impairment affects only fu, but to
`different degrees by altering RF50 from
`0.5 to 50 1/h,
`renal impairment affects only CLint' but to
`different degrees by altering Sint from 30
`to 900, and
`renal impairment affects both fu and CLint
`by setting RF50 at 1 1/h and Sint at 300.
`
`These three cases described a scenario
`where the CRF affects PK of a dmg with high
`
`Boehringer Ex. 2020
`Mylan v. Boehringer Ingelheim
`IPR2016-01565
`Page 5
`
`

`
`Yuan and Venitz
`
`248
`
`' Table 1.
`
`' Scenarios
`
`Lo~ PPB/Low Efl,
`Low PPB/High ER
`High PPB/High ER
`
`PPB and low ER, such as drug X. Additional
`simulations were conducted for three other
`scenarios, i.e. , for drugs with pharrnaco(cid:173)
`kinetic characteristics different from drug X
`with respect to plasma protein binding (PPB)
`and hepatic extraction ratio (ER), by setting
`the respective parameters as in Table I.
`
`Results
`
`The total plasma clearance of drug X
`(CL101) , calculated from equation 1, did not
`appear to be correlated with CLcrea (Figure
`
`Fig.1a
`
`Fig. 1c
`
`Ia). However, when the total plasma clear(cid:173)
`ance was corrected for PPB, the unbound
`clearance of drug X (CL101") seemed to in(cid:173)
`crease with the increasing CLcrca (Figure. I b),
`with a slope that is significantly different
`from zero (r2 = 0.34, p = 0.004). Since most of
`the drug X is eliminated by hepatic metabo(cid:173)
`lism, its unbound plasma clearance reflects
`the unbound intrinsic clearance. Thus, the ac(cid:173)
`tual unbound intrinsic clearance, CLint> was
`calculated from equations 3 and 4. A 2.7-fold
`difference in its mean value was found be(cid:173)
`tween group I and IV. When plotting the CLint
`as a continuous function ofCLcrw a linear re(cid:173)
`lationship became apparent (r2 = 0.32, p =
`0.006, Figure lc). Using equation 6, the inter(cid:173)
`cept (CLint0) was estimated as 2000 1/h, and
`the slope (Sim) as 308. This means that for
`equation 7, CLintN was 5000 1/h, and the slope
`S 'int was 30. In contrast to the relationship be(cid:173)
`tween CLint and CLcrea• an inverse hyperbolic
`relationship was observed for the relationship
`between fu and CLcrea· Protein binding of drug
`
`• •
`•
`
`• • •
`
`•
`
`+---- f - -- - t - - - - +---+----~---~
`4
`10
`12
`
`Cl.,.., (lJhr)
`
`• •
`
`•
`
`6
`Cl.,., (lJhr)
`
`10
`
`12
`
`Fig~re. 1a:. Hepatic clearance of drug X uncorrected for PPB (Clheptot) versus creatinine clearance
`(Clcrea); b: unbound hepatic clearance of drug X (Clhepu) versus creatinine clearance (Clcrea); c: unbound
`intrinsic clearance of drug X (CL;01 ) versus creatinine clearance (Clcrea); Symbols and the line represents the
`observed and predicted values, respectively; Clinto =- 2000 1/h and Sint = 300; d: unbound fraction of drug X
`(fu) versus creatinine clearance (Clcreal; symbols and the line represents the observed and predicted val(cid:173)
`ues, respectively; fu max = 0.006; fu min= 0.002 and RF50 = 1 1/h.
`
`Boehringer Ex. 2020
`Mylan v. Boehringer Ingelheim
`IPR2016-01565
`Page 6
`
`

`
`PK of highly metabolized drugs in renally impaired patients
`
`249
`
`14 T
`13 +
`12 +
`
`-+- RF50=0.5
`-RF50=2.5
`........-RF50=5
`~RF50=50
`
`CL.c, •• (Lihr)
`
`Figure 2. Simulated total hepatic clearance of
`drug X (Clheptot) with renal disease affecting only
`PPB.
`
`25
`
`I
`- =;..._ Sint=30
`-Sint=1ru0
`__,._ Sint=200
`
`1
`
`~Sint=300
`
`1
`
`c~.a .. (llhr)
`
`Figure 3. Simulated total hepatic clearance of
`drug X (Clheptot) with renal disease affecting only
`Clint·
`
`14
`
`12
`
`~ 10
`g
`1 8
`(3
`
`6
`
`4
`
`0
`
`2
`
`4
`
`6
`
`8
`
`C Lcrea (Lihr)
`
`Figure 4. Scenario 1: Simulated hepatic clearance
`of low ER drug with high PPB.
`
`X was not affected by chronic renal disease
`until CLcrea declined to approximately 3 l/h.
`When renal function decreased further, a dra(cid:173)
`matic increase in the unbound fraction of the
`drug occurred. Modeling this relationship
`showed that the creatinine clearance at which
`the unbound fraction increased by 50% was 1
`1/h (- 20 ml/min), and the minimal and maxi(cid:173)
`mum fraction of unbound drug was 0.2% and
`0.6%, respectively (r2 = 0.29, Figure ld). It is
`noteworthy that these disease-PK models for
`fu vs. CLcrea and CLint vs. CLcrea were parsi(cid:173)
`monious in the choice and the number of pa(cid:173)
`rameters, therefore, the significant residual
`variability remained as seen in the low values
`ofr2•
`Subsequent PK simulations demonstrate
`that if CRF modifies only one factor, namely
`fu or CLinh the hepatic clearance follows the
`pattern of that particular factor (Figure 2, 3).
`The higher the RF 50 value, the weaker is the
`influence of renal disease on fu, the less is the
`impact of CRF on hepatic clearance (Figure
`2). As for Sint• the larger the value, the more
`prominent is the effect ofCRF on CLint and on
`hepatic clearance (Figure 3). When the renal
`disease exerts its influence on both factors si(cid:173)
`multaneously, as is the case for drug X, the
`overall change in the observed hepatic clear(cid:173)
`ance is the combination of CRF effect on each
`individual factor. Under this circumstance,~
`biphasic U-shaped hepatic clearance (hence,
`CLtot for a highly metabolized drug) versus
`CLcrea relationship may occur (Figure 4).
`For drugs with lowER and low PPB, renal
`function will not significantly influence PPB
`since it is already low. As a result, the hepatic
`clearance, Clheptot, of such a drug (Figure 5)
`is determined by the effect of CRF on un(cid:173)
`bound intrinsic clearance as simulated in Fig(cid:173)
`ure 3. The higher the degree of influence that
`CRF casts on intrinsic clearance of such a
`drug, the more of a difference will be seen in
`its hepatic clearance between the NRF and
`CRF groups. However, the CRF influence on
`hepatic clearance also depends on the actual
`magnitude of hepatic clearance. With the
`same degree of CRF influence on intrinsic
`clearance, i.e., the same Sint value, the change
`in hepatic clearance of drugs with high ER
`will be less discemable than that with lowER.
`For example, when Sint is set at 300 (assuming
`that PPB does not change in either cases), the
`change in hepatic clearance for a high ER
`
`Boehringer Ex. 2020
`Mylan v. Boehringer Ingelheim
`IPR2016-01565
`Page 7
`
`

`
`Yuan and Venitz
`
`250
`
`30
`
`28
`
`~ 26
`s ...
`J 24
`0
`
`22
`
`20+---------4---------~---------+-------~
`0
`2
`4
`6
`8
`
`45
`
`43
`
`~ 41
`li! 39
`..J
`0
`
`37
`
`. C'-crea (Uhr)
`
`Figure 5. Scenario II: Simulated hepatic clearance
`of low ER drug with low PPB.
`
`drug is less than 10% (Figure 6), as compared
`
`2
`
`4
`C'-crea (Uhr)
`
`6
`
`8
`
`Figure 6. Scenario Ill: Simulated hepatic clear(cid:173)
`ance of high ER drug with low PPB.
`
`to the 50% change simulated for lowER drug
`(Figure 3). Ifboth PPB and intrinsic clearance
`change dramatically in the CRF patients for a
`highER drug, the pattern ofCRF influence on
`its hepatic clearance (hence, total plasma
`clearance of a highly metabolized drug) will
`be the same as that of! owER drug. However,
`the magnitude of such a change between NRF
`and CRF groups may be too small to be ob(cid:173)
`served clinically (Figure 7).
`
`Discussion
`
`In this study, we have demonstrated that
`CRF can affect both intrinsic hepatic meta-
`
`bolic clearance and protein binding of a drug,
`thereby influencing the disposition even of
`highly metabolized drugs. Because the over(cid:173)
`all influence ofCRF may not result in a linear
`relationship between CLhep with CLcrea> the
`severe group may not necessarily represent
`the worst case scenario in terms of the impact
`ofCRF on hepatic clearance. Our analysis for
`drug X showed that protein binding of this
`drug was affected by an inverse hyperbolic
`relationship with creatinine clearance. Since
`the inflection point of this relationship was at
`11/h (i.e., the mean creatinine clearance in the
`severe group IV), fu was affected most promi(cid:173)
`nently by severe renal impairment. Since in(cid:173)
`creased fu tends to lead to a higher total clear(cid:173)
`ance, a higher plasma clearance would be
`expected. However, this trend was counter(cid:173)
`acted by the progressive decline in CLint with
`CRF. As a result, the severely renally im(cid:173)
`paired group did not show the lowest hepatic
`clearance value for drug X, but the moderate
`group (II) did.
`
`The fact that CRF may cause an increased
`free fraction for acidic drugs is a well(cid:173)
`documented phenomenon. Our simulations
`demonstrate that if CRF exerts its influence
`only on fu, the total plasma clearance of a
`highly metabolized drug, CL10t, would actu(cid:173)
`ally increase with decreasing renal function,
`but the magnitude of the disease impact is de(cid:173)
`pendent on the parameter RF 50 (Figure 2).
`When the RF 50 is small for a lowER drug, the
`effect of CRF on CLhcptot,
`thus
`total
`steady-state plasma concentrations, can be
`dramatic. The study on meloxicam, a highly
`plasma protein-bound drug that is completely
`absorbed and predominantly metabolized, il(cid:173)
`lustrates this point [Turck eta!. 1996]: in pa(cid:173)
`tients with end-stage renal failure, the total
`plasma concentration of meloxicam de(cid:173)
`creased by 60% as the result of increased total
`clearance. But the free concentration of the
`drug in CRF patients remained the same as
`that in NRF subjects, indicating that unbound
`intrinsic clearance, CLint> of the drug stayed
`unaltered in renal patients while the total
`clearance, CLtot' increased secondary to the
`decreased protein binding. Unfortunately, the
`disposition of meloxicam in the mild and
`moderate renal impairment groups were not
`investigated in the same study, therefore, the
`relationship between fu and CLcrea for this
`drug (RF 50) can not be detennined. The pa-
`
`Boehringer Ex. 2020
`Mylan v. Boehringer Ingelheim
`IPR2016-01565
`Page 8
`
`

`
`35 1
`
`33
`
`-.:-
`~ 31
`ga.
`J 29
`
`(.)
`
`27
`
`25
`
`0
`
`PK of highly metabolized drugs in renally impaired patients
`
`251
`
`way has a different Sint• thus showing differ(cid:173)
`ent susceptibility to renal disease. However,
`alternatively, metabolism in hepatocyte may
`not be
`the
`culprit,
`but
`rather
`the
`hepato-cellular uptake mechanisms as far as
`the effect of chronic uremia is concerned.
`
`Being a lowER drug, it is not surprising to
`see that plasma clearance of dmg X was a
`composite of changes in fu and CLint· The in(cid:173)
`travenous route of administration for this
`dmg allowed us to directly estimate systemic
`clearance without having to take oral absorp(cid:173)
`tion into consideration. After an intravenous
`administration of clorazepate, a prodmg for
`the highly metabolized drug desmethyl(cid:173)
`diazepam, Ochs et al. [1984] reported no
`change
`in systemic clearance of total
`desmethyldiazepam in severe renal patients
`as compared to normal subjects. But the frac(cid:173)
`tion of free desmethyldiazepam (fu) increased
`in plasma and the free hepatic clearance
`(CLhep ") decreased by more than 200% in the
`severely impaired renal patients. As with
`drug X, desmethyldiazepam is highly pro(cid:173)
`tein-bound and 99% of the drug is metabo(cid:173)
`lized. However, since we did not have the data
`from the intermediate groups, we could not
`estimate the shape of desmethyldiazepam
`clearance curve with respect to creatinine
`clearance. But the biphasic relationship be(cid:173)
`tween the plasma clearance and renal func(cid:173)
`tion derived from drug X was observed for
`cerivastatin [Vormfelde et al. 1999], a drug
`shares the similar pharmacokinetic character(cid:173)
`istics with drug X. Cerivastatin has oral clear(cid:173)
`ance of 13 Uh and PPB of99.5%. The drug is
`I 00% absorbed. Even though metabolism ac(cid:173)
`counts for 100% of its elimination, the total
`clearance of cerivastatin is elevated signifi(cid:173)
`cantly in CRF patients compared to the values
`in NRF subjects [Vormfelde eta!. 1999]. The
`most dramatic change was observed not in the
`severe group, but in the moderate group. As
`our simulations indicate, the shape and mag(cid:173)
`nitude change in plasma clearance for this
`group of dmgs depends on the magnitude of
`renal influence on each factor of fu or CLint·
`Unlike the simulated cases 1 and 2 for the low
`ER drugs, the most dramatic change in case 3
`may not be in the severe group. If this is the
`case,
`simply
`studying
`the
`drug
`pharmacokinetics in severe group and normal
`group may overlook the change in disposition
`for the intermediate groups.
`
`2
`
`4
`
`6
`
`8
`
`Clcrea (Llhr)
`
`Figure 7. Scenario IV: Simulated hepatic clear(cid:173)
`ance of high ER drug with high PPB.
`
`rameter RF 50 determines the susceptibility of
`fu to CRF progression. When RF50 increases
`gradually, CLheprot may seem to be in a in(cid:173)
`verse linear relationship with CLcrea• given
`that CLint is not affected by renal function
`(Figure 2).
`
`According to our simulations in case 2 for
`dmg X, renal
`impairment may decrease
`hepatic clearance, thus reducing systemic
`clearance of a highly metabolized dmg, by
`decreasing CLint of a low ER dmg without af(cid:173)
`fecting fu (Figures 3, 5). This scenario is espe(cid:173)
`cially applicable for drugs with low protein
`binding, because renal impairment should not
`significantly affect protein binding that is al(cid:173)
`ready low. Oxycodone is a low extraction
`drug with plasma protein binding of 38%. In
`healthy volunteers, more than 90% of the
`drug is eliminated by hepatic metabolism.
`Kirvela et al. [ 1996] studied pharmaco(cid:173)
`kinetics of oxycodone in normal subjects and
`uremic patients undergoing renal
`trans(cid:173)
`platation. The authors reported significantly
`higher plasma concentrations of oxycodone
`in the uremic group than in the control group,
`suggesting an effect of CRF on its intrinsic
`metabolic clearance. The impact of renal dis(cid:173)
`ease on non-renal biotransformation depends
`on the slope parameter, Sin~> according to our
`simulations. It has been proposed in the litera(cid:173)
`ture that the effect of renal disease on hepatic
`clearance may depend on specific hepatic
`metabolic pathways [Gibson 1986, Touchette
`and Slaughter 1991 ]. Therefore, it is possible
`that each different biotransformation path-
`
`Boehringer Ex. 2020
`Mylan v. Boehringer Ingelheim
`IPR2016-01565
`Page 9
`
`

`
`Yuan and Venitz
`
`Our simulations, furthennore, show that
`CRF impacts the systemic disposition of high
`ER drugs to a much les