Longitudinal assessment of everolimus in
`de novo renal transplant recipients over the
`first post-transplant year: Pharmacokinetics,
`exposure-response relationships, and
`influence on cyclosporine
`
`Objective: Our objective was to characterize the steady-state pharmacokinetics of everolimus and
`cyclosporine (INN, ciclosporin) when coadministered in de novo kidney allograft recipients during the
`first year after transplantation.
`Methods: This study was a multicenter randomized double-blind study of 101 patients who were randomly
`assigned 1:1:1 to receive everolimus tablets at doses of 0.5 mg, 1 mg, or 2 mg twice daily with cyclosporine
`and prednisone. Blood sampling for the pharmacokinetics of everolimus and cyclosporine was performed
`on day 1, on weeks 1, 2, 3, and 4, and on months 2, 3, 6, 9, and 12. Everolimus dose-proportionality
`and stability over time were assessed in the context of linear regression and ANOVA models. Everolimus
`exposure-response relationships between area under the blood concentration-time curve (AUC) and
`changes in platelets, leukocytes, and lipids were explored with the median-effect model. Potential differ-
`ences in cyclosporine dosing and pharmacokinetics at different levels of everolimus exposure were assessed
`in the context of ANOVA.
`Results: Everolimus steady state was reached on or before day 7, with a median 3-fold accumulation of drug
`exposure compared with that after the first postoperative dose. Both steady-state maximum concentration
`and AUC were dose proportional over the full dose range when assessed on day 1, as well as for the full
`duration of the study at steady state. There was evidence for longitudinal stability in AUC of everolimus
`during the course of the study. The interindividual pharmacokinetic variability for AUC was 85.4% and
`intraindividual, interoccasion variability was 40.8%. Age (range, 17-69 years), weight (range, 49-106 kg),
`and sex (65 men and 36 women) were not significant contributors to variability. There was an increasing
`incidence of transient thrombocytopenia (≤100 × 109/L) with increasing everolimus AUC (P = .03).
`Cyclosporine doses, trough concentrations, and AUC exhibited similar temporal patterns during the course
`of the study regardless of the co-administered everolimus dose level (P = .13, .82, and .76, respectively).
`Conclusions: Everolimus exhibited dose-proportional, stable exposure during the first post-transplant year.
`For a 4-fold range of everolimus doses there were no differential effects on cyclosporine dosing or phar-
`macokinetics. (Clin Pharmacol Ther 2001;69:48-56.)
`
`John M. Kovarik, PhD, Barry D. Kahan, MD, Bruce Kaplan, MD, Marc Lorber, MD,
`Michael Winkler, MD, Marisel Rouilly, PhD, Christophe Gerbeau, PhD,
`Natalie Cambon, MD, Robert Boger, MD, and Christiane Rordorf, MD,
`on behalf of the Everolimus Phase 2 Study Group* Basel, Switzerland, Houston, Texas,
`Livingston and East Hanover, NJ, New Haven, Conn, and Hannover, Germany
`
`From Novartis Pharmaceuticals, Basel and East Hanover; University
`of Texas Medical School-Houston; St Barnabas Medical Center,
`Livingston; Yale University School of Medicine, New Haven; and
`Medizinische Hochschule, Hannover.
`*Study group members: S. Cockfield, University of Alberta Hospi-
`tals, Edmonton; D. Hricik, University Hospitals of Cleveland,
`Ohio; K. Rigg and M. Shehata, Nottingham City Hospital, Eng-
`land; W. Land, Klinikum Grosshadern, Munich, Germany; and E.
`Mueller, A. G. Schmidt, S. Wehr, M. M. Wilkie, and R. L. Wong,
`Novartis Pharmaceuticals, East Hanover, NJ.
`
`Supported by research grants from the National Institute of Diabetes,
`Digestive and Kidney Diseases (NIDDK#38016-12) and Novartis
`Pharmaceuticals.
`Received for publication Aug 9, 2000; accepted Nov 14, 2000.
`Reprint requests: John Kovarik, PhD, Novartis Pharma AG, Build-
`ing WSJ 27.4093, 4002 Basel, Switzerland.
`Copyright © 2001 by Mosby, Inc.
`0009-9236/2001/$35.00 + 0 13/1/112969
`doi:10.1067/mcp.2001.112969
`
`48
`
`Ex. 1050-0001
`
`

`
`CLINICAL PHARMACOLOGY & THERAPEUTICS
`VOLUME 69, NUMBER 1
`
`Kovarik et al 49
`
`Temporal changes in the pharmacokinetics of
`immunosuppressants used in acute rejection prophylaxis
`after organ transplantation have been reported.1-4 Two
`examples are cyclosporine (INN, ciclosporin) and
`mycophenolate mofetil. For the non-microemulsion for-
`mulation of cyclosporine the relationship between dose
`and area under the blood concentration-time curve
`(AUC) changes in the first 3 to 4 months after a trans-
`plant and stabilizes thereafter.1,2 Although this pattern is
`also evident for the microemulsion formulation, the
`period of dose-AUC stability is achieved earlier (by the
`end of the first month) compared with the non-
`microemulsion formulation.3 This suggests that the oral
`absorption and bioavailability of cyclosporine in the
`early post-transplant period (which is improved with the
`microemulsion formulation) is probably contributing to
`this phenomenon. For mycophenolate mofetil, dose-
`normalized AUC is lower in the early post-transplant
`period and increases slowly over several months by an
`average 50% to reach late post-transplant values. Because
`bioavailability is nearly complete throughout this period,
`poor absorption cannot explain these observations. Spec-
`ulation has centered on changes in protein binding and
`enterohepatic recirculation.4 These examples underscore
`the importance of characterizing the longitudinal phar-
`macokinetics of new immunosuppressants, especially in
`the early post-transplant period so that time-dependent
`changes in the disposition of the immunosuppressants
`may be taken into account in the dosing regimen to yield
`more stable systemic exposure over time.
`Everolimus (development name, RAD) is a macro-
`cyclic lactone immunosuppressant that is primarily
`metabolized and eliminated in the bile. At therapeutic
`concentrations, more than 75% of everolimus is parti-
`tioned into red blood cells and approximately 75% of
`the plasma fraction is protein bound. Everolimus blocks
`growth factor–driven transduction signals in the T-cell
`response to alloantigen5 and thus acts at a later stage
`than the calcineurin inhibitors cyclosporine and
`tacrolimus. The complimentary modes of action of
`everolimus and cyclosporine suggest a synergistic inter-
`action that has indeed been shown in vitro and in vivo
`in preclinical models.5 This provided a rationale for the
`addition of everolimus to cyclosporine-based immuno-
`suppression. Subsequent phase 1 studies showed good
`tolerability and an acceptable side-effect profile when
`everolimus was administered for 1 month to stable
`patients with renal transplants who were receiving a
`cyclosporine-prednisone regimen.6,7 A recently com-
`pleted phase 2 study constituted the first experience
`with everolimus in patients immediately after kidney
`transplantation in the de novo setting.8 In that trial,
`
`extensive pharmacokinetic evaluations were performed
`during a 1-year period after transplantation for longitu-
`dinal characterization of the disposition of everolimus.
`The clinical results will be reported elsewhere; this arti-
`cle focuses on the pharmacokinetic results of the study
`in which dose proportionality, longitudinal stability in
`exposure, and pharmacokinetic variability with poten-
`tial covariates were addressed.
`The inhibitory effects of everolimus are not restricted
`to T cells. Everolumus also inhibits the signals provided
`by some hematopoietic and nonhematopoietic cell
`growth factors.5 In this context, phase 1 experience with
`this compound in patients indicated that treatment with
`everolimus may be associated with decreases in
`platelets and leukocytes and increases in lipids that gen-
`erally occur in the first 2 months after initiation of treat-
`ment.6,7 This has also been the experience with
`sirolimus.9,10 Because we used a broader range of
`everolimus doses in this phase 2 study than are cur-
`rently under investigation in phase 3 efficacy trials, it
`provided the opportunity for us to screen for exposure-
`response relationships between AUC and changes in
`laboratory parameters.
`Both everolimus and cyclosporine are extensively
`biotransformed by means of CYP3A and are substrates
`for P-glycoprotein. Therefore there is a potential for a
`drug-drug interaction when these two agents are co-
`administered. Although everolimus was the primary
`focus of this study,
`the longitudinal influence of
`everolimus on cyclosporine pharmacokinetics was also
`explored as a secondary objective.
`
`METHODS
`Study design and dose regimens. This was a 1-year,
`randomized, double-blind trial performed at 8 study cen-
`ters. The study protocol was approved by local medical
`ethics committees for each center. One hundred three
`de novo kidney allograft recipients gave written
`informed consent to participate in the study and were
`randomly assigned to receive everolimus (Certican,
`Novartis Pharmaceuticals, East Hanover, NJ) at oral
`doses of 0.5 mg (n = 34), 1 mg (n = 34), or 2 mg (n =
`35) twice daily in addition to cyclosporine and pred-
`nisone. The first dose of everolimus was given once it
`had been ascertained that the allograft was functional
`(maximum, 48 hours after transplantation). This was
`defined as study day 1. Everolimus was supplied as a
`tablet formulation in strengths of 0.25 mg and 1 mg with
`matching placebos. Patients were instructed to take the
`prescribed number of tablets every 12 hours simultane-
`ously with cyclosporine. Cyclosporine (Neoral, Novar-
`tis Pharmaceuticals) was initiated orally at 6 to 12
`
`Ex. 1050-0002
`
`

`
`50 Kovarik et al
`
`CLINICAL PHARMACOLOGY & THERAPEUTICS
`JANUARY 2001
`
`mg/kg/d in two divided doses. Thereafter, the doses
`were adjusted to maintain whole blood predose concen-
`trations in the range of 150 to 400 ng/mL in the first
`post-transplant month and in the range of 5 to 300
`ng/mL from month 2 to month 12. Prednisone was
`dosed according to a protocol-specified taper.
`Pharmacokinetic assessments. Blood sampling for
`the pharmacokinetics of everolimus and cyclosporine
`was performed at protocol-scheduled visits throughout
`the study duration: day 1, weeks 1, 2, 3, and 4, and
`months 2, 3, 6, 9, and 12. At each visit, either a full
`pharmacokinetic profile that consisted of 10 blood sam-
`ples (predose and 0.5, 1, 1.5, 2, 3, 4, 6, 9, and 12 hours)
`or an abbreviated profile (predose, and 1, 2, 5, and 8
`hours or predose, and 1, 2, and 5 hours) was performed
`during the morning dosing interval. At the week 3 and
`month 2 and 12 visits, only a predose trough sample
`was obtained. For the 5-point abbreviated profiles, the
`12-hour concentration was estimated by log-linear
`decrease from the 5-hour and 8-hour measured concen-
`trations; for the 4-point abbreviated profiles, the 12-
`hour concentration was assumed to be the same as the
`predose measured concentration under the assumption
`of steady state. When maximum concentration (Cmax)
`and AUC from the full profiles in this study were re-
`evaluated on the basis of the 4- or 5-sample abbrevi-
`ated approach described previously, they were highly
`correlated (r2 = 0.91 for Cmax and r2 ≥ 0.96 for AUC);
`this indicated that parameters from full and abbreviated
`sampling could be pooled for analysis.
`Bioanalysis of everolimus. Everolimus whole-blood
`concentrations were determined with a validated
`enzyme-linked immunosorbent assay. Performance was
`assessed on the basis of a 5-point quality control con-
`centration range from 2 to 80 ng/mL of everolimus.
`Coefficients of variation ranged from 13.3% to 16.1%
`and bias ranged from –7.0% to –1.8%. The assay quan-
`tification limit was 2 ng/mL.
`Sample preparation consisted of the mixing of 250 µL
`of blood sample with 1 mL of extraction buffer (Tris
`buffer, 0.05 mol/L, pH 9) and 1 mL of extraction solvent
`(diethyl ether with 0.04% Tween 20). The samples were
`shaken for 10 minutes and centrifuged for 10 minutes at
`4000 rpm and 4°C. They were subsequently frozen in
`dry ice, and the organic phase was decanted and evapo-
`rated at 40°C for 20 to 30 minutes. The dry residues were
`reconstituted with 120 µL of a reconstitution solution
`that consisted of 200 ng/mL mouse antirapamycin anti-
`body in phosphate-buffered saline solution (PBS) buffer
`(1:10 dilution) with 1% bovine serum albumin. For the
`determination of nonspecific binding, the dry residue
`was dissolved in dilution buffer without antibody. The
`
`samples were shaken for 10 minutes and then incubated
`for at least 15 minutes at room temperature.
`Plastic 96-well assay plates were coated overnight
`with 120 µL per well of 10 µg/mL goat anti-mouse anti-
`body in PBS buffer (Goat Anti-Mouse IgG Fc,
`Immunopure, Rockford, Ill) at a temperature of approx-
`imately 4°C. The plates were then washed with wash-
`ing buffer (PBS buffer with 0.05% Tween 20) and
`blocked with 300 µL per well of blocking solution by
`incubation for 5 minutes at room temperature. The
`plates were washed and 100 µL per well of the
`processed blood samples was added. After 1 hour of
`incubation at room temperature with gentle shaking, 20
`µL of tracer solution (19 ng/mL of rapamycin
`biotinyled in PBS buffer with 1% bovine serum albu-
`min and 0.04% Tween 20) were added to each well. The
`plates were incubated overnight at approximately 4°C
`and then washed with PBS buffer with 0.05% Tween 20
`and incubated with 120 µL per well of streptavidin per-
`oxidase solution (diluted 1:10,000 in assay buffer) for
`15 minutes in the dark at room temperature with gentle
`shaking. Plates were washed with assay buffer and 120
`µL per well of o-phenylenediamine dihydrochloride
`solution (10 mg in 50 mL of 0.05 mol/L citrate/phos-
`phate buffer and 20 µL of hydrogen peroxide 30%) was
`added, and then the plates were shaken gently for 5 to
`6 minutes at room temperature. The reaction was
`stopped with 50 µL per well of 2 N sulfuric acid, and
`the optical density was measured at 492/620 nm.
`Bioanalysis of cyclosporine. Cyclosporine whole
`blood concentrations were determined with the use of
`the Incstar CYCLO-Trac radioimmunoassay kit (Still-
`water, Minn) according to the manufacturer’s directions.
`Performance was assessed on the basis of a 5-point qual-
`ity control concentration range from 30 to 1250 ng/mL.
`Coefficients of variation ranged from 5.1% to 10.6%
`and bias ranged from –12.0% to 0.4%. The assay quan-
`tification limit was 30 ng/mL.
`Pharmacokinetic and statistical evaluation. Con-
`ventional noncompartmental steady-state pharmaco-
`kinetic parameters were derived for both everolimus
`and cyclosporine, including the predose trough concen-
`tration (Cmin); the Cmax; the time to reach maximum
`concentration (tmax); the AUC during a dosing interval,
`by trapezoidal summation; the average concentration
`(Cavg), calculated as the quotient of AUC/12; and the
`peak-trough fluctuation, calculated as (Cmax – Cmin)/
`Cavg. The accumulation ratio was derived from the ratio
`of AUCs from day 7 and day 1.
`Dose proportionality for the Cmax and AUC of
`everolimus was assessed on day 1 (first dose) and day
`7 (steady state) with linear regression analysis of the
`
`Ex. 1050-0003
`
`

`
`CLINICAL PHARMACOLOGY & THERAPEUTICS
`VOLUME 69, NUMBER 1
`
`Kovarik et al 51
`
`Table I. Everolimus pharmacokinetic parameters*
`0.5 mg bid
`
`First dose
`tmax (h)
`Cmax (ng/mL)
`AUC (ng · h/mL)
`Steady state†
`Cmin (ng/mL)
`tmax (h)
`Cmax (ng/mL)
`Cmax/dose (ng/mL/mg)
`AUC (ng · h/mL)
`AUC/dose (ng · h/mL/mg)
`Cavg (ng/mL)
`PTF (%)
`
`3 (1-12)
`2.0 ± 2.1
`8 ± 12
`
`1.5 ± 1.8
`2 (1-5)
`5.0 ± 2.9
`10.0 ± 5.8
`34 ± 23
`68 ± 46
`2.8 ± 1.9
`84 ± 38
`
`bid, Twice a day; PTF, peak-trough fluctuation.
`*Values are median (range) for tmax and mean ± standard deviation for all other parameters.
`†Steady-state data are from week 1.
`
`1 mg bid
`
`3 (2-9)
`5.6 ± 3.7
`28 ± 23
`
`4.7 ± 2.6
`2 (1-5)
`11.6 ± 4.4
`11.6 ± 4.4
`81 ± 34
`81 ± 34
`6.7 ± 2.8
`97 ± 38
`
`2 mg bid
`
`3 (2-12)
`9.8 ± 7.0
`56 ± 37
`
`9.5 ± 5.2
`2 (1-8)
`21.9 ± 10.5
`11.0 ± 5.3
`164 ± 78
`82 ± 39
`13.6 ± 6.5
`90 ± 47
`
`parameter versus dose. Steady-state, dose-normalized
`parameters from day 7 to the end of the study were also
`evaluated in an ANOVA model with dose, subject-
`within-dose, visit, and dose-by-visit interaction terms.
`In the absence of a dose-by-visit interaction, lack of a
`dose-effect was taken as support for dose proportional-
`ity during the full study course; lack of a visit effect
`was evidence for longitudinal stability in exposure.
`Steady-state AUC per dose from the replicate pharma-
`cokinetic profiles was assessed with a two-way ANOVA
`with subject and visit as sources of variation. The mean
`square from the subject term was taken as a measure of
`interindividual variance, and the mean square from the
`error term was taken as a measure of intraindividual,
`interoccasion variance. The corresponding coefficients
`of variation were calculated as the standard deviation
`(square root of the mean square) divided by the grand
`mean of the parameter. The contribution of conven-
`tional demographic covariates to the pharmacokinetic
`variability was explored with graphical and regression
`techniques for continuous variables (age and weight)
`and with unpaired two-sided t tests for categorical vari-
`ables (sex and ethnicity).
`Dose-normalized cyclosporine pharmacokinetic param-
`eters were evaluated with a global ANOVA model identi-
`cal to that for everolimus in which the dose term referred
`to the everolimus dose level. The absence of a dose-by-
`visit interaction and of a dose-effect was taken as support
`that the influence of everolimus on cyclosporine was not
`different among the everolimus dose levels.
`Exposure-response relationships. Relationships
`between everolimus steady-state AUC and the incidence
`of thrombocytopenia (≤100 × 109/L), leukopenia (≤4.0
`
`× 109/L), hypertriglyceridemia (>2.9 mmol/L), and
`hypercholesterolemia (>6.5 mmol/L) were explored
`with the median-effect principle.11 This model relates
`the fraction of the population affected (fa) and unaf-
`fected (fu = 1 – fa) with respect to a given laboratory
`parameter change on the one hand and to the drug expo-
`sure (AUC) and the exposure at which half the popula-
`tion is affected (median effect [AUCm]) on the other
`hand: fa/fu = (AUC/AUCm)m. The relationship is lin-
`earized on the logarithmic scale as follows: log(fa/fu)
`= m log(AUC) – m log(AUCm). In this relationship, m
`is a Hill-type coefficient that describes the sigmoidic-
`ity in the exposure-response relationship. The average
`AUC for each patient was determined from the week 1,
`2, and 4 profiles and the overall distribution divided
`into quartiles. The fraction of patients whose labora-
`tory parameter did (fa) and did not (fu) exceed the
`defined cutoff value listed previously in the first two
`post-transplant months was determined in each of the
`quartiles. The resulting log(fa/fu) versus log(AUC) rela-
`tionship was assessed by linear regression analysis.
`Goodness of fit of the model to the data was indicated
`by a regression coefficient (r value) of more than 0.8;
`a regression P value of less than .05 indicated a signif-
`icant relationship between exposure and the occurrence
`of a laboratory parameter change.
`
`RESULTS
`Demographics and patient disposition. Of the 103
`patients enrolled in the study, 101 provided at least one
`pharmacokinetic profile and were included in this analy-
`sis. The two unevaluable patients withdrew from the trial
`in the first week and did not undergo pharmacokinetic
`
`Ex. 1050-0004
`
`

`
`52 Kovarik et al
`
`CLINICAL PHARMACOLOGY & THERAPEUTICS
`JANUARY 2001
`
`A
`
`B
`
`Fig 1. Everolimus mean AUC trajectories during 9 months after transplants (A) and everolimus
`trough concentrations (Cmin) to 12 months after transplants (B) at dose levels of 0.5 mg (solid
`squares), 1 mg (solid circles), and 2 mg twice a day (solid triangles). Bars designate 95% confi-
`dence intervals.
`
`blood sampling. There were 65 men and 36 women who
`were 44.4 ± 11.6 years old and who weighed 76.7 ± 15.4
`kg. The majority of the patients were white (n = 82);
`there were 9 black patients, 1 Asian patient, and 9
`patients of other ethnicities.
`Everolimus pharmacokinetic profiles were evaluable
`at each visit if the patient remained on the dose to which
`he or she was randomly assigned and had not missed
`doses within 5 days before the visit. Over the 1-year
`duration of the study, 18 patients had dose reductions for
`safety reasons, 31 patients had transient dose interrup-
`tions, and 16 patients were removed from the study (8
`unsatisfactory therapeutic response, 3 died, 3 withdrew
`
`consent, and 2 violated the study protocol). Conse-
`quently, the number of evaluable profiles at each of the
`steady-state visits were: 94 (week 1), 88 (week 2), 83
`(week 4), 71 (month 3), 58 (month 6), and 55 (month 9).
`First-dose everolimus pharmacokinetics. A total of
`95 profiles were obtained after the first dose of
`everolimus on day 1: 33, 32, and 30 profiles, respec-
`tively, were obtained from patients who were receiving
`0.5, 1, and 2 mg twice daily. First-dose pharmacokinetic
`parameters are summarized in Table I. In postoperative
`conditions, there was generally a 1-hour lag time until
`everolimus was quantifiable in blood, and the peak con-
`centration occurred at 3 hours after the dose. Both Cmax
`
`Ex. 1050-0005
`
`

`
`CLINICAL PHARMACOLOGY & THERAPEUTICS
`VOLUME 69, NUMBER 1
`
`Kovarik et al 53
`
`Fig 2. Median-effect plots of everolimus steady-state AUC and incidence of hypercholesterolemia
`(solid diamonds), hypertriglyceridemia (solid triangles), and thrombocytopenia (solid circles) in
`patients with renal transplants during first 2 post-transplant months. For these laboratory parame-
`ters, the slopes were 1.27, 0.66, and 0.70, and the correlation coefficients were 0.94, 0.83, and 0.97,
`respectively. Incidence is expressed as the fraction of the population affected and unaffected (fa/fu).
`
`and AUC rose proportionally with increasing dose (r =
`0.57, P < .001 for Cmax and r = 0.61, P < .001 for AUC).
`Steady-state everolimus pharmacokinetics. Steady
`state was reached on or before day 7 on the basis of the
`ANOVA. Between the first dose and steady state there
`was a median accumulation ratio of 2.9. A total of 449
`steady-state profiles were evaluable between week 1
`and month 9 (100 with full 12-hour sampling and 349
`with abbreviated sampling). Steady-state parameters
`from week 1 are summarized in Table I. Cmin increased
`in a dose-proportional manner during week 1 at steady
`state (r = 0.47, P < .001). The Cmax was subsequently
`reached by 2 hours after the dose and also increased in
`a dose-proportional manner (r = 0.51; P < .001). The
`peak trough fluctuation was relatively narrow and
`approximately 90% at all dose levels. Overall steady-
`state exposure (AUC) exhibited dose proportionality
`during week 1 (r = 0.53, P < .001). In addition, the
`absence of dose-effects in the global ANOVA on dose-
`normalized Cmax (P = .65) and AUC (P = .10) from the
`week 1 to month 9 profiles indicated that steady-state
`dose proportionality was maintained during the full
`study course.
`Everolimus longitudinal stability and pharmacoki-
`netic variability. Everolimus mean AUC trajectories are
`shown in Fig 1, A for all three dose levels. There was
`no dose-by-visit interaction (P = .85); this indicated
`that the pattern in AUCs over time was similar for all
`dose levels. A significant visit effect was noted (P =
`
`.004); subsequent pairwise contrasts attributed this to a
`lower than average pooled AUC per dose at week 2 (68
`ng · h/mL) and a higher than average pooled value at
`month 9 (91 ng · h/mL) relative to the surrounding vis-
`its (70-80 ng · h/mL). Therefore the temporal pattern
`was generally stable, with two relatively small devia-
`tions at isolated visits, during the full study course.
`The global longitudinal estimate was 85.4% for
`interindividual variability for AUC per dose and was
`40.8% for intraindividual,
`interoccasion variability.
`Explorations for covariates that contributed to interindi-
`vidual variability in exposure indicated that age, which
`ranged from 17 to 69 years, explained only 2.6% of the
`variability and weight, which ranged from 49 to 106 kg,
`contributed less than 1.0% on the basis of the coefficients
`of determination (r2) from linear regression analysis.
`Likewise, AUC per dose was not different between sexes
`(P = .39) or ethnicities (white versus nonwhite, P = .30).
`Everolimus serial trough concentrations. A total of
`651 steady-state trough concentrations were available
`between week 1 and month 12. There was no dose-by-
`visit interaction (P = .98); this confirmed the impres-
`sion that the Cmin trajectories were similar among the
`everolimus dose levels over time (Fig 1, B). At the low-
`est dose level, several troughs were below the assay
`quantification limit and were set to 0 ng/mL by conven-
`tion. This biased the mean values downward and
`resulted in a significant dose-effect (P = .03). Specifi-
`cally, Cmin per dose pooled across visits at the upper two
`
`Ex. 1050-0006
`
`

`
`54 Kovarik et al
`
`CLINICAL PHARMACOLOGY & THERAPEUTICS
`JANUARY 2001
`
`A
`
`B
`
`Fig 3. Cyclosporine (CsA) morning doses (A) and cyclosporine mean AUC per dose trajectories to
`month 9 (B) in the everolimus 0.5-mg (solid squares), 1-mg (solid circles), and 2-mg twice a day (solid
`triangles) dose cohorts during 1 year after transplants. Bars designate 95% confidence intervals.
`
`dose levels was consistent with dose proportionality (4.5
`± 3.0 ng/mL/mg at 1 mg and 4.8 ± 2.4 ng/mL/mg at 2
`mg) but was less than proportional at the 0.5-mg dose
`level (3.2 ± 3.7 ng/mL/mg). A significant visit effect (P
`< .001) was detected; subsequent pairwise comparisons
`attributed the visit effect to an isolated occasion of low
`values at week 3 relative to the surrounding visits dur-
`ing the study course. Steady-state Cmin was well corre-
`lated with AUC during the year-long study; this yielded
`a correlation coefficient (r value) of 0.88.
`Everolimus exposure-response relationships. The
`average minimum platelet count over the first 2 post-
`transplant months was 160 ± 51 × 109/L; on average, the
`minimum count occurred by day 16. As shown in Fig 2,
`there was a significant relationship between everolimus
`
`AUC and the incidence of thrombocytopenia (P = .03).
`The average maximum triglyceride level was 3.8 ± 1.8
`mmol/L (337 ± 159 mg/dL) and the maximum choles-
`terol level was 7.8 ± 1.4 mmol/L (302 ± 54 mg/dL),
`which occurred on average by days 29 and 35, respec-
`tively. There were trends toward an increased incidence
`of both hypertriglyceridemia (P = .14) and hypercholes-
`terolemia (P = .06) with increasing everolimus AUC. The
`average minimum leukocyte count was 5.7 ± 1.7 × 109/L,
`and it occurred by day 29. The incidence of leukopenia
`was not related to everolimus exposure (P = .35).
`Cyclosporine dosing. Dosing began at 240 ± 74 mg,
`252 ± 50 mg, and 245 ± 60 mg twice a day in the
`ascending everolimus dose cohorts. As shown in Fig 3,
`A, cyclosporine doses were reduced over time (P <
`
`Ex. 1050-0007
`
`

`
`CLINICAL PHARMACOLOGY & THERAPEUTICS
`VOLUME 69, NUMBER 1
`
`Kovarik et al 55
`
`.001) as is conventional in the first post-transplant year.
`Cyclosporine doses did not differ among the everolimus
`dose groups (P = .13) nor did the temporal patterns dif-
`fer over the study course as shown by lack of a dose-
`by-visit interaction (P = .99).
`Cyclosporine pharmacokinetics. A total of 586
`cyclosporine trough concentrations were evaluable during
`the year-long study course. Trough concentrations were
`significantly reduced during the first post-transplant year
`(P < .001). The temporal patterns did not differ among
`everolimus cohorts (P = .58 for dose-by-visit interaction)
`nor were cyclosporine troughs different among the
`everolimus dose levels (P = .82). This was also true for
`dose-normalized troughs (P = .70 and .25, respectively).
`There was a total of 417 cyclosporine profiles (109
`from full 12-hour sampling and 308 from abbreviated
`sampling). Because cyclosporine doses were reduced
`over time, the corresponding AUCs differed signifi-
`cantly across study visits (P < .001). The temporal pat-
`tern for AUC per dose shown in Fig 3, B exhibited an
`increase during the first month after the transplant with
`stability thereafter despite the fact that the cyclosporine
`dose was being continuously decreased. This was the
`result of an alteration in the dose-exposure relationship
`during the first month as has been described in previ-
`ously published series with cyclosporine (Neoral) in the
`absence of everolimus.3 There was no significant dose-
`by-visit interaction (P = .94) or everolimus dose-effect
`(P = .76) for AUC or the corresponding dose-normal-
`ized values; this confirmed that the temporal patterns
`in cyclosporine AUC were similar regardless of
`everolimus dose cohort.
`
`DISCUSSION
`This study provided the first pharmacokinetic and
`exposure-response data for everolimus in patients with
`renal transplants longitudinally followed up from the
`de novo period in the first 3 post-transplant months
`through the metastable phase from months 3 to 6 and
`into the stable phase from months 6 to 12. The impor-
`tance of careful assessment of the disposition of an
`immunosuppressant during this time frame stems from
`the experience with cyclosporine and mycophenolate
`mofetil. For both agents there are changes in the dose-
`AUC relationship in the first 3 months after a trans-
`plant; these changes may need to be taken into account
`in dosing.3,4 The cyclosporine data from this study
`clearly show this phenomenon, as illustrated in Fig 3.
`The pharmacokinetics of everolimus after the first
`dose indicated some delays in absorption (median tmax
`of 3 hours) that were probably related to postsurgical
`gastrointestinal ileus rather than to the dosage formu-
`
`lation because there was rapid absorption (tmax of 1-2
`hours) by the next assessment (day 7). A good absorp-
`tion profile was maintained during the trial period and
`was similar to that characterized in stable patients with
`renal transplants in a previous study.6 Although absorp-
`tion was delayed after the first dose, both Cmax and
`AUC were dose proportional; this was also the case lon-
`gitudinally at steady state over the full study course.
`The range of doses used in this trial extended below
`and above those currently being assessed in multicen-
`ter efficacy studies. Therefore the dose proportionality
`observed in this study implies that safety- or efficacy-
`related dose titrations within the potential therapeutic
`dose range should yield proportional changes in AUC
`in the clinical setting. Although it is currently unknown
`whether dose titration will need to be guided by thera-
`peutic drug monitoring as is used for some other
`immunosuppressants, it was encouraging to observe the
`good correlation between predose trough concentra-
`tions and AUC, which suggests that monitoring of
`trough concentrations may serve as a marker for over-
`all exposure if needed.
`With regard to temporal stability in exposure, visual
`inspection of the AUC and Cmin trajectories as shown
`in Fig 1 did not show any specific longitudinal trends.
`Although there were significant visit effects in ANOVA,
`further pairwise comparisons of the parameters across
`visits indicated that values at a few isolated nonsequen-
`tial visits were either higher or lower than at the other
`visits. These transient fluctuations may have been
`caused by the different number of evaluable profiles at
`each visit because of patient withdrawals, dose inter-
`ruptions, and samples not being obtained. Nonetheless,
`the magnitudes of these isolated deviations in exposure
`are unlikely to be of clinical relevance, and a generally
`stable pattern in everolimus exposure was observed.
`Pharmacokinetic variability has conventionally been
`assessed in a replicate single-dose study with a rela-
`tively short washout period between assessments so that
`the only source of variability is theoretically intraindi-
`vidual, interoccasion effects. In this study, pharmaco-
`kinetic profiling was performed under outpatient dos-
`ing conditions during an extended period of 9 months
`(that is, from day 7 to month 9) during which the
`patient’s clinical condition changed. Both the long
`period of assessment and the change in clinical status
`must be kept in mind when one consideres the intrain-
`dividual, interoccasion coefficient of variation of 40.8%
`for dose-normalized AUC. This contrasts with the esti-
`mate from a previous study of everolimus in clinically
`stable patients with renal transplants who underwent
`pharmacokinetic profiling on two occasions 1 week
`
`Ex. 1050-0008
`
`

`
`56 Kovarik et al
`
`CLINICAL PHARMACOLOGY & THERAPEUTICS
`JANUARY 2001
`
`apart and in whom the coefficient of intraindividual,
`interoccasion variability was 10% to 19%.7 Post-trans-
`plant changes in renal function in our study population
`is unlikely to have contributed to variability because
`renal clearance of everolimus is minimal. The coeffi-
`cient of interindividual variation in the present study
`was 85.4%