- Email: [email protected]
Reduced Inter- and Intraindividual Variability in Cyclosporine Pharmacokinetics from a Microemulsion Formulation
Reduced Inter- and Intraindividual Variability in Cyclosporine Pharmacoklnetics from a Microemulsion Formulation JOHN
M. KOVARM'~, EDGARA.
MUELLER',
JOHANNES 6. VAN
BREEt, WOLFGANG TETZLOFF*, AND
KLAUSKUTZ'
Received March 15, 1993, from the 'Department of Human Pharmacology, Drug Safety Assessment, Sandoz Pharma. Ltd., 4002 Basle, Switzerland, the %iopharmaceutics Department, Sandoz Pharma. Ltd.. Bash Switzerknd, and the giphar CRF Institute Accepted for publication July 19, 1993'. for Clinical Pharmacology, Hohenkirchen-Siegertsbrunn, Germany. Abstract 0 The inter- and intraindividual variability of cyclosporine
pharmacokinetics from a microemulsion formulation were compared with the currently marketed formulation in a sequential bioreplication study. Twenty-four heatthy male volunteers were randomizedto receive each formulation on two separate occasions; the referencetreatment was a single oral dose of 300 mg of Sandimmuneand the test treatment was a single oral dose of 180 mg of Sandimmune Neoral, both given as soft gelatin capsules. Serial venous blood samples were obtained over a period of 48 h after each administration, and cyclosporine concentrationswere measured in whole blood by a specificmonoclonal R I A method. Between-and within-subject variabilitieswere quantified from the appropriate sums of squares from analysis of variance and statistlcally compared between formulations. Both inter- and intraindividual variation for the peak concentration,time to reach the peak, area under the curve, and terminal half-lifeof the test formulation were significantlyreduced (p < 0.05) with two exceptions. For area under the curve between subjects (p < 0.2) and peak concentration within subjects (p < O.l), trends toward reduced variabiltty for the test formulation were evident. These results were further reflected in the inter-and intraindividualcoefficientsof variation of the pharmacokinetic parameters that ranged from 3 to 22% for the test formulation comparedwtth 19to41%forthereferenceformulation. Incomparison with the currently marketed formulation, reduced variability in the pharmacokinetics of cyclosporine following oral administration of Sandimmune Neoral provides a more predictable and consistent concentration-time profile. Cyclosporinehas become an established immunosuppressant in the management of organ transplant patients. However, following oral administration, its absolute bioavailability and pharmacokinetics are markedly variable.' A major source of variability appears to be due to poor absorption.2 Contributing factors include the relatively high molecular weight and very high lipophilicityof cyclosporine,the existence of an absorption window in the small intestine: and poor solubility in aqueous fluid^.^ Release of the drug from the currently marketed formulation may also influence this variability. When cyclosporine is administered as Sandimmune, a crude oil-in-water emulsion forms in aqueous gastrointestinal fluids with cyclosporine mainly distributed in lipid droplets. These droplets must be dispersed by bile and pancreatin to form mixed micelles for subsequent drug absorption.5~6 In an attempt to optimize cyclosporine oral absorption, a new oral formulation (Sandimmune Neoral) was developed that incorporates the drug in a microemulsion preconcentrate. Following oral administration, it immediately forms a microemulsion in aqueous fluids, simulating a mixed micellar phase. Cyclosporine is therefore rapidly availablefor absorption so that the entire length of the absorption window can be utilized. In circumventing the lipid dispersion step, absorption is also less influenced by bile flow7 and pancreatin. It was anticipated that improving the oral absorption of cyclosporine would have an impact on its pharmacokinetic variability. To assess this, a sequential bioreplication study was undertaken in healthy 0
Abstract published in Adoance ACS Abstracts, January 1, 1994.
444 /Journal of pharmaceutical scknces Vol. 83, No. 3, h4arch 1994
volunteers to characterize the inter- and intraindividual variability of cyclosporine pharmacokinetics following oral administration of the microemulsion formulation in comparison with the currently marketed formulation.
Experimental Section Subjects-Twenty-four healthy male volunteers aged 24 7 years (mean f SD) and weighing 72 f 8 kg completed the study. All participants signed a written informed consent after they had been informed of the nature and details of the study. Volunteers were evaluated for general good health on the basis of medical history, physical examination, electrocardiogram, routine biochemical and hematologic tests, and hepatitis and HIV serologies. None of the subjects was concurrently taking other medications. Study Design-The study protocol was approved by a local medical ethics committee and was performed in accordance with the Declaration of Helsinki and with current European Community and US.Food and Drug Administration guidelines for Good Clinical Practice. The investigation was conducted as a sequential bioreplication studye in which subjects were randomized in a crossover allocation to receive each formulation on two separate occasions in the sequence RRTT or TTFtR, where R is the reference treatment andT is the testtreatment. Successive administrations were separated by a washout phase of 2 weeks. The R treatment consisted of a single 300-mg oral dose of Sandimmune (as three 100-mgsoft gelatin capsules; lot no. Y 144 0890, Sandoz Pharma, Ltd.) and the T treatment consisted of a single 180-mg oral dose of Sandimmune Neoral (asthree 60-mgsoft gelatin capsules; lot no. X 177 0991, Sandoz Pharma, Ltd.). The milligram doses were chosen based on the results of a previous comparative bioavailability study in healthy fasting subjects to provide comparable cyclosporineexposure (area under the curve) from both formulations. Subjects were confined to the study center from 14 h before until 48 h after each drug administration. On the days of drug dosing, they fasted from 12 h before administration until 4 h after. The doses were taken with 200 mL of tap water and, for a given subject, were administered at the same time of day on each dosing occasion. Beginning 4 h after administration, all subjects were given standardized, scheduled meals that were identical on all dosing days. Additionally, fluid intake was standardized and no alcohol was allowed 2 days perior to or during the period of confinement. Subjects were not allowed to remain in a supine position for the f i i t 4h after dosing. Followingeach drug administration, venous blood samples for the determination of cyclosporine in whole blood were obtained predose and then 0.25,0.5,0.75,1,1.5,2,2.5,3,3.5, 4,4.5,5,6,8,10,12,16,20,24,28,32,36,40,and 48 h after administration. Samples were collected in EDTA-containing tubes, gently inverted several times, and frozen at -20 O C . BioanalyticalMethods-Concentrations of cyclosporineAinwhole blood were assayed with the commercially available Sandimmune radioimmunoassay kit (Sandoz Ltd., Basle, Switzerland) that is based on the use of a monoclonal antibody specific for the parent comp0und.g At quality control concentrations of 6.25, 12.5,100, 400, and 1600 ng/ mL, the respective accuracies were 32.1, 8.5, -2.4, -1.7, and 3.0%;the intraassay coefficients of variation were 41.1,16.8,5.4,4.1, and 5.7; and the interassay coefficients of variation were 20.1, 10.8, 12.4, 8.9, and 11.3%. The overall detection limit of the assay (5.6 f 0.8 ng/mL, n = 25) was calculated from the mean concentration corresponding to 95% binding. The limit of quantification (intra-assay coefficient of variation, 130%) was set at 12.5 ng/mL. Pharmacokinetic Evaluation-Noncompartmental pharmacokinetic characteristics were derived by standard methods with concen-
0022-3549/94/ 120O-444$04.50/0
@ 1994. Amerlcan ChemlCaI SOCkV 8nd American pharmaceutical A s 0 ~ 4 t l o n
Table 1-Pharmacoklnetlc Characteristlcs of Cyclorporlne (Mean f SD) Following Two Replicate Oral Admlnlstratlons of 300mg Reference Formulations and 180-mg Ted Formulations In 24 Healthy Volunteers Reference Formulation
Pharmacokinetic Characteristic
Administration 1
2.17 f 0.82 802 f 284 3496 f 901 11.7 f 4.7
tmaxv h
ng/mL A N b , h.ng/mL t1121 h &.b#
Test Formulation
Administration 2
Administration 1
Administration 2
2.27 f 0.97 761 f 249 3212 f 958 8.4 f 4.0
1.44 f 0.40 1037 f 178 3407 f 692 7.8 f 1.7
1.54 f 0.53 983 f 196 3441 f 633 7.3 f 1.6
trations above the assay limit of quantification (12.5 ng/mL). The maximum whole blood (b) concentration C-b and the time of its Occurrence t- were compiled from the concentration-time data. The area under the curve (Aucb) was calculated by the h e a r trapezoidal rule to the last blood concentration C(t& above the limit of quantification and extrapolated to infinity by the addition of the term &&/X., where C(t& and X. are the predicted concentration at time t . and the terminal dispositionrate constant, respectively. The rate constant w a determined ~ by nonlinear regression of the terminal portion of the concentrationtime profile with all concentrations above the assay limit of detection weighted as the reciprocal of the aasay variance. The extrapolations contributed, on the average, 5.9 f 3.5% (n = 96) to the total AUCb. Statistical Evaluation-Comparison of Formulations-The primary pharmacokinetic characteristics for characterizing the rate and extent of cyclosporine absorption from the immediate-release formulations were t,, c-b (rate), and AUCb (extent).lo These parameters were globallyassessed by analysisof variance (ANOVA) after ascertaining normality of distribution (Wilk-Shapiro test) and homogeneity of variances (Levine test). Subsequent multiple comparisons testing (Newman-Keuls) demonstrated no significant differences in these parameters between replicate administrations for each formulation (for all tests, the statistical power to detect a difference of f20% around the reference mean was >a%). Data were therefore pooled by taking the mean for the two replicate administrations of each subject. The formulations were compared by subjecting the pooled parameters to ANOVA with power testing. Between- and Within-Subject Variation-For each formulation, the ANOVA mean s u m of squares associated with between subject (MSb) and within subject (MS,) variabilities of the primary pharmacokinetic characteristics were derived from the replicate administrations (i.e., R1 and R2 for the reference formulation and T1 and T2 for the test formulation) as follows: MSb by pooling the mean squares for subjects and subjects per group and MS, by pooling the variation due to periods and treatments with the residual.8 The MS for each pharmacokinetic characteristic were subsequently compared between the two formulations by an F-ratio test to determine whether the variabilities were significantly different? As an additional meaeureof variability, the percent coefficient of variation (CV) was derived from the respective ANOVA mean squares as follows:
0
6
12
18
24
30
=
a
40
Time [h] Figure 1-Interindividuei comparisonof cyclosporineconcentration-time profllesfdlowlngsingleoraiadmlnishrrtion of 300-mgreferenceformuiation to 24 volunteers. Inset shows the initial portionof the profile on a Iinearlinear scale.
".0
\
W ''
42
36
i
100 grand mean
In eqs 1and 2, cvband CV, are the coefficients of variation associated with between and within subject variability, R is the number of replicate administrations, and grand mean is the mean of all observations (n = 48). Statistical hypotheses were tested at the 0.05 significance level. All statistical calculationswere performed with the RS/1 software package.ll
I
0
6
12
18
24
30
36
42
48
Time [h]
Results and Discusslon
Figure2-InterlndivMual comparison of cyclosporlneconcentration-time profiles following single oral administration of 180-mg test formulation to 24 volunteers. Inset shows the initial porHon of the profile on a iinearlinear scale.
Clinical Observations-The tolerabilityof both cyclosporine formulations was good. There were no clinically relevant or drug-related changes in vital signs, electrocardiograms, or biochemical and hematologic tests. Comparison of Formulations-Cyclosporine pharmacokineticcharacteristicsare summarizedin Table 1 for each replicate
administration. As anticipated, the rate of cyclosporine absorption differed between formulations as indicated by comparison of data pooled from the two replicate administrations for each formulation. t , was earlier [1.49 f 0.41 versus 2.22 f 0.82 h (p < 0.00l)l and C-b was moderately higher [1010 f 163versus 781 f 234n g / d (p < 0.00l)lfor the test formulation. Jownal of pharmaceutlcel Scknces / 445 Vol. 83,No. 3,Merch 1994
Table 2-Inter- and Intrasubject Varlabllltys In the Pharmacoklnetic Characteristics Of Cyclosporlne from Two Oral Formulations In 24 Healthy Volunteers Intersubject Variability Characteristic tmax C-.b AUCb ti12
Reference MSb Test M S b 1.35 109, 926 1, 332, 735 26.42
0.33 53,402 785, 451 2.78
Intrasubject Variability
F-Ratio Reference (Significance) cvb, % 4.09 (p C 0.01) 2.06 (p C 0.05) 1.70 (p < 0.2) 9.50 (p < 0.01)
33.5 25.2 20.2 22.2
Test
Reference
cvb, %
MS,
22.4 13.3 17.2 3.0
0.24 32, 281 419, 515 16.52
Test MS,
F-Ratio Reference Test (Significance) CV,, % CV,, %
0.11 2.18 (p C 0.05) 17, 533 1.84 (p < 0.1) 91, 371 4.59 (p C 0.01) 6.19 (p C 0.01) 2.67
22.3 23.0 19.3 40.6
22.2 13.1 8.8 21.6
MSb and MS, are the ANOVA mean sum of squares associated with between and within individual variability; cvb and CV, are the between and within subject percent coefficients of variation. t m a x [hl
.
[ng h/mL]
AUC
Cmax [ng/mL]
I 15001 6000
I
formulation with the exception of Cm,b for which a strong trend (p < 0.1) toward lower variability was evident. Intrasubject CVs for the primary parameters of cyclosporine ranged from 9 to 22% for the test formulation compared with 19 to 41 % for the reference formulation. It should be noted that in the case of t, the difference identified by comparison of the ANOVA MS values is obscured when comparing the CVs because these measures of variability are derived by further calculation from the MS values and can therefore be influenced by differences in the grand means (see eq 2).
Conclusions lO0Ol
I " R T O R T
R
T
Formulation
Figure 3-Intraindividual comparison of the primary pharmacokinetic characteristics of cyclosporine following replicate oral administrations of 300-mg reference (R) formulation and 180-mg test (T) formulation to 24 healthy volunteers.
The doses for each formulation were preselected to provide the same systemic availability of cyclosporine; this was confirmed by comparison of the AUCbs that were comparable (p = 0.57; statistical power 100%) for the test (3424 f 627 h.ng/mL) and the reference (3354 f 816 ha ng/mL) formulations. The terminal disposition half-life was statistically different on the two assessment occasions for the reference formulation (p < 0.01). A between-formulation comparison could not be performed inasmuch as the reference data could not be pooled. Intersubject Variability-Synoptic views of the cyclosporine whole blood concentration-time profiles for the first of the replicate administrations of each formulation are illustrated graphically in Figures 1 and 2. As shown in Table 2, intersubject variability as assessed by ANOVA MSb was significantly less for the pharmacokinetic parameters of the test formulation with the exception of AUCb; for the latter parameter, a trend (p < 0.2) toward lower variability was evident for the test formulation. Intersubject CVs for the primary parameters of cyclosporine ranged from 3 to 22% for the test formulation compared with 20 to 34% for the reference formulation. Intrasubject Variability-Intrasubject variability in the rate and extent of absorption is represented in Figure 3. As compiled in Table 2, comparison of t h e ANOVA MS, demonstrated that the variability within subjects was significantly less for the pharmacokinetic parameters of the test
446 /Journal of Pharmaceutical Sciences Vol. 83, No. 3, M r c h 1994
Sandimmune Neoral is a microemulsion formulation of cyclosporine that represents an improvement in the delivery of the drug for oral absorption. Altering the oral formulation had a demonstrable impact on pharmacokinetic variability. In comparison with the currently marketed formulation, reduced inter- and intraindividual variability in cyclosporine pharmacokinetics from Sandimmune Neoral yields a more consistent and predictable concentration-time profile. It is anticipated that this will translate clinically into greater facility in reaching the desired target blood Concentration range, fewer maintenance dosage adjustments, and perhaps a reduction in the frequency of concentration monitoring necessary in routine patient management.
References and Notes 1. Ptachinski, R. J.; Venkataramanan, R.; Burckart, G. J. Clin. Pharmacokinet. 1986,11, 107-132. 2. Grevel, J. Transplant R o c . 1986, ld(supp1. 5 ) , 9-15. 3. Drewe, J.; Beglinger, C.; Kissel, T. Br. J. Clin. Pharmacol. 1992, 33, 39-43. 4. Reymond, J. P.;Steimer,J. L.; Niederberger,W. J.Pharmacokinet. Biopharm. 1989,16, 331-353. 5. Reymond, J.; Sucker, H. Pharm. Res. 1988,5,673-676. 6. Reymond, *J.; Sucker, H.; Vonderscher, J. Pharm. Res. 1988, 5, 677-679. 7. Trull, A. K.; Tan, K. K. C.; Uttridge, J.; Bauer, T.; Alexander, G. J. M.; Jamieson, N. V. Lancet 1993,341,433. 8. Albert, K. S.; Smith, R. B. In Drug Absorption and Disposition: Statistical Considerations; Albert, K. S., Ed.; American Pharmaceutical Association: Washington, D.C., 1980; pp 87-113. 9. Ball, P. E.; Munzer, H.; Keller, H. P.; Abisch, E.; Rosenthaler, J. Clin. Chem. 1988,34,257-260. 10. Steinijans, V. W.; Hauschke, D. Znt. J. Clin. Pharmacol. Ther. Toxicol. 1990,28, 105-110. 11. RS/1 for VAX and Micro VAX. Release 4.1 software:Bolt Beranek
and Newman Inc. (BBN) Software Products Corporation: Cambridge, MA, 1989.
M. KOVARM'~, EDGARA.
MUELLER',
JOHANNES 6. VAN
BREEt, WOLFGANG TETZLOFF*, AND
KLAUSKUTZ'
Received March 15, 1993, from the 'Department of Human Pharmacology, Drug Safety Assessment, Sandoz Pharma. Ltd., 4002 Basle, Switzerland, the %iopharmaceutics Department, Sandoz Pharma. Ltd.. Bash Switzerknd, and the giphar CRF Institute Accepted for publication July 19, 1993'. for Clinical Pharmacology, Hohenkirchen-Siegertsbrunn, Germany. Abstract 0 The inter- and intraindividual variability of cyclosporine
pharmacokinetics from a microemulsion formulation were compared with the currently marketed formulation in a sequential bioreplication study. Twenty-four heatthy male volunteers were randomizedto receive each formulation on two separate occasions; the referencetreatment was a single oral dose of 300 mg of Sandimmuneand the test treatment was a single oral dose of 180 mg of Sandimmune Neoral, both given as soft gelatin capsules. Serial venous blood samples were obtained over a period of 48 h after each administration, and cyclosporine concentrationswere measured in whole blood by a specificmonoclonal R I A method. Between-and within-subject variabilitieswere quantified from the appropriate sums of squares from analysis of variance and statistlcally compared between formulations. Both inter- and intraindividual variation for the peak concentration,time to reach the peak, area under the curve, and terminal half-lifeof the test formulation were significantlyreduced (p < 0.05) with two exceptions. For area under the curve between subjects (p < 0.2) and peak concentration within subjects (p < O.l), trends toward reduced variabiltty for the test formulation were evident. These results were further reflected in the inter-and intraindividualcoefficientsof variation of the pharmacokinetic parameters that ranged from 3 to 22% for the test formulation comparedwtth 19to41%forthereferenceformulation. Incomparison with the currently marketed formulation, reduced variability in the pharmacokinetics of cyclosporine following oral administration of Sandimmune Neoral provides a more predictable and consistent concentration-time profile. Cyclosporinehas become an established immunosuppressant in the management of organ transplant patients. However, following oral administration, its absolute bioavailability and pharmacokinetics are markedly variable.' A major source of variability appears to be due to poor absorption.2 Contributing factors include the relatively high molecular weight and very high lipophilicityof cyclosporine,the existence of an absorption window in the small intestine: and poor solubility in aqueous fluid^.^ Release of the drug from the currently marketed formulation may also influence this variability. When cyclosporine is administered as Sandimmune, a crude oil-in-water emulsion forms in aqueous gastrointestinal fluids with cyclosporine mainly distributed in lipid droplets. These droplets must be dispersed by bile and pancreatin to form mixed micelles for subsequent drug absorption.5~6 In an attempt to optimize cyclosporine oral absorption, a new oral formulation (Sandimmune Neoral) was developed that incorporates the drug in a microemulsion preconcentrate. Following oral administration, it immediately forms a microemulsion in aqueous fluids, simulating a mixed micellar phase. Cyclosporine is therefore rapidly availablefor absorption so that the entire length of the absorption window can be utilized. In circumventing the lipid dispersion step, absorption is also less influenced by bile flow7 and pancreatin. It was anticipated that improving the oral absorption of cyclosporine would have an impact on its pharmacokinetic variability. To assess this, a sequential bioreplication study was undertaken in healthy 0
Abstract published in Adoance ACS Abstracts, January 1, 1994.
444 /Journal of pharmaceutical scknces Vol. 83, No. 3, h4arch 1994
volunteers to characterize the inter- and intraindividual variability of cyclosporine pharmacokinetics following oral administration of the microemulsion formulation in comparison with the currently marketed formulation.
Experimental Section Subjects-Twenty-four healthy male volunteers aged 24 7 years (mean f SD) and weighing 72 f 8 kg completed the study. All participants signed a written informed consent after they had been informed of the nature and details of the study. Volunteers were evaluated for general good health on the basis of medical history, physical examination, electrocardiogram, routine biochemical and hematologic tests, and hepatitis and HIV serologies. None of the subjects was concurrently taking other medications. Study Design-The study protocol was approved by a local medical ethics committee and was performed in accordance with the Declaration of Helsinki and with current European Community and US.Food and Drug Administration guidelines for Good Clinical Practice. The investigation was conducted as a sequential bioreplication studye in which subjects were randomized in a crossover allocation to receive each formulation on two separate occasions in the sequence RRTT or TTFtR, where R is the reference treatment andT is the testtreatment. Successive administrations were separated by a washout phase of 2 weeks. The R treatment consisted of a single 300-mg oral dose of Sandimmune (as three 100-mgsoft gelatin capsules; lot no. Y 144 0890, Sandoz Pharma, Ltd.) and the T treatment consisted of a single 180-mg oral dose of Sandimmune Neoral (asthree 60-mgsoft gelatin capsules; lot no. X 177 0991, Sandoz Pharma, Ltd.). The milligram doses were chosen based on the results of a previous comparative bioavailability study in healthy fasting subjects to provide comparable cyclosporineexposure (area under the curve) from both formulations. Subjects were confined to the study center from 14 h before until 48 h after each drug administration. On the days of drug dosing, they fasted from 12 h before administration until 4 h after. The doses were taken with 200 mL of tap water and, for a given subject, were administered at the same time of day on each dosing occasion. Beginning 4 h after administration, all subjects were given standardized, scheduled meals that were identical on all dosing days. Additionally, fluid intake was standardized and no alcohol was allowed 2 days perior to or during the period of confinement. Subjects were not allowed to remain in a supine position for the f i i t 4h after dosing. Followingeach drug administration, venous blood samples for the determination of cyclosporine in whole blood were obtained predose and then 0.25,0.5,0.75,1,1.5,2,2.5,3,3.5, 4,4.5,5,6,8,10,12,16,20,24,28,32,36,40,and 48 h after administration. Samples were collected in EDTA-containing tubes, gently inverted several times, and frozen at -20 O C . BioanalyticalMethods-Concentrations of cyclosporineAinwhole blood were assayed with the commercially available Sandimmune radioimmunoassay kit (Sandoz Ltd., Basle, Switzerland) that is based on the use of a monoclonal antibody specific for the parent comp0und.g At quality control concentrations of 6.25, 12.5,100, 400, and 1600 ng/ mL, the respective accuracies were 32.1, 8.5, -2.4, -1.7, and 3.0%;the intraassay coefficients of variation were 41.1,16.8,5.4,4.1, and 5.7; and the interassay coefficients of variation were 20.1, 10.8, 12.4, 8.9, and 11.3%. The overall detection limit of the assay (5.6 f 0.8 ng/mL, n = 25) was calculated from the mean concentration corresponding to 95% binding. The limit of quantification (intra-assay coefficient of variation, 130%) was set at 12.5 ng/mL. Pharmacokinetic Evaluation-Noncompartmental pharmacokinetic characteristics were derived by standard methods with concen-
0022-3549/94/ 120O-444$04.50/0
@ 1994. Amerlcan ChemlCaI SOCkV 8nd American pharmaceutical A s 0 ~ 4 t l o n
Table 1-Pharmacoklnetlc Characteristlcs of Cyclorporlne (Mean f SD) Following Two Replicate Oral Admlnlstratlons of 300mg Reference Formulations and 180-mg Ted Formulations In 24 Healthy Volunteers Reference Formulation
Pharmacokinetic Characteristic
Administration 1
2.17 f 0.82 802 f 284 3496 f 901 11.7 f 4.7
tmaxv h
ng/mL A N b , h.ng/mL t1121 h &.b#
Test Formulation
Administration 2
Administration 1
Administration 2
2.27 f 0.97 761 f 249 3212 f 958 8.4 f 4.0
1.44 f 0.40 1037 f 178 3407 f 692 7.8 f 1.7
1.54 f 0.53 983 f 196 3441 f 633 7.3 f 1.6
trations above the assay limit of quantification (12.5 ng/mL). The maximum whole blood (b) concentration C-b and the time of its Occurrence t- were compiled from the concentration-time data. The area under the curve (Aucb) was calculated by the h e a r trapezoidal rule to the last blood concentration C(t& above the limit of quantification and extrapolated to infinity by the addition of the term &&/X., where C(t& and X. are the predicted concentration at time t . and the terminal dispositionrate constant, respectively. The rate constant w a determined ~ by nonlinear regression of the terminal portion of the concentrationtime profile with all concentrations above the assay limit of detection weighted as the reciprocal of the aasay variance. The extrapolations contributed, on the average, 5.9 f 3.5% (n = 96) to the total AUCb. Statistical Evaluation-Comparison of Formulations-The primary pharmacokinetic characteristics for characterizing the rate and extent of cyclosporine absorption from the immediate-release formulations were t,, c-b (rate), and AUCb (extent).lo These parameters were globallyassessed by analysisof variance (ANOVA) after ascertaining normality of distribution (Wilk-Shapiro test) and homogeneity of variances (Levine test). Subsequent multiple comparisons testing (Newman-Keuls) demonstrated no significant differences in these parameters between replicate administrations for each formulation (for all tests, the statistical power to detect a difference of f20% around the reference mean was >a%). Data were therefore pooled by taking the mean for the two replicate administrations of each subject. The formulations were compared by subjecting the pooled parameters to ANOVA with power testing. Between- and Within-Subject Variation-For each formulation, the ANOVA mean s u m of squares associated with between subject (MSb) and within subject (MS,) variabilities of the primary pharmacokinetic characteristics were derived from the replicate administrations (i.e., R1 and R2 for the reference formulation and T1 and T2 for the test formulation) as follows: MSb by pooling the mean squares for subjects and subjects per group and MS, by pooling the variation due to periods and treatments with the residual.8 The MS for each pharmacokinetic characteristic were subsequently compared between the two formulations by an F-ratio test to determine whether the variabilities were significantly different? As an additional meaeureof variability, the percent coefficient of variation (CV) was derived from the respective ANOVA mean squares as follows:
0
6
12
18
24
30
=
a
40
Time [h] Figure 1-Interindividuei comparisonof cyclosporineconcentration-time profllesfdlowlngsingleoraiadmlnishrrtion of 300-mgreferenceformuiation to 24 volunteers. Inset shows the initial portionof the profile on a Iinearlinear scale.
".0
\
W ''
42
36
i
100 grand mean
In eqs 1and 2, cvband CV, are the coefficients of variation associated with between and within subject variability, R is the number of replicate administrations, and grand mean is the mean of all observations (n = 48). Statistical hypotheses were tested at the 0.05 significance level. All statistical calculationswere performed with the RS/1 software package.ll
I
0
6
12
18
24
30
36
42
48
Time [h]
Results and Discusslon
Figure2-InterlndivMual comparison of cyclosporlneconcentration-time profiles following single oral administration of 180-mg test formulation to 24 volunteers. Inset shows the initial porHon of the profile on a iinearlinear scale.
Clinical Observations-The tolerabilityof both cyclosporine formulations was good. There were no clinically relevant or drug-related changes in vital signs, electrocardiograms, or biochemical and hematologic tests. Comparison of Formulations-Cyclosporine pharmacokineticcharacteristicsare summarizedin Table 1 for each replicate
administration. As anticipated, the rate of cyclosporine absorption differed between formulations as indicated by comparison of data pooled from the two replicate administrations for each formulation. t , was earlier [1.49 f 0.41 versus 2.22 f 0.82 h (p < 0.00l)l and C-b was moderately higher [1010 f 163versus 781 f 234n g / d (p < 0.00l)lfor the test formulation. Jownal of pharmaceutlcel Scknces / 445 Vol. 83,No. 3,Merch 1994
Table 2-Inter- and Intrasubject Varlabllltys In the Pharmacoklnetic Characteristics Of Cyclosporlne from Two Oral Formulations In 24 Healthy Volunteers Intersubject Variability Characteristic tmax C-.b AUCb ti12
Reference MSb Test M S b 1.35 109, 926 1, 332, 735 26.42
0.33 53,402 785, 451 2.78
Intrasubject Variability
F-Ratio Reference (Significance) cvb, % 4.09 (p C 0.01) 2.06 (p C 0.05) 1.70 (p < 0.2) 9.50 (p < 0.01)
33.5 25.2 20.2 22.2
Test
Reference
cvb, %
MS,
22.4 13.3 17.2 3.0
0.24 32, 281 419, 515 16.52
Test MS,
F-Ratio Reference Test (Significance) CV,, % CV,, %
0.11 2.18 (p C 0.05) 17, 533 1.84 (p < 0.1) 91, 371 4.59 (p C 0.01) 6.19 (p C 0.01) 2.67
22.3 23.0 19.3 40.6
22.2 13.1 8.8 21.6
MSb and MS, are the ANOVA mean sum of squares associated with between and within individual variability; cvb and CV, are the between and within subject percent coefficients of variation. t m a x [hl
.
[ng h/mL]
AUC
Cmax [ng/mL]
I 15001 6000
I
formulation with the exception of Cm,b for which a strong trend (p < 0.1) toward lower variability was evident. Intrasubject CVs for the primary parameters of cyclosporine ranged from 9 to 22% for the test formulation compared with 19 to 41 % for the reference formulation. It should be noted that in the case of t, the difference identified by comparison of the ANOVA MS values is obscured when comparing the CVs because these measures of variability are derived by further calculation from the MS values and can therefore be influenced by differences in the grand means (see eq 2).
Conclusions lO0Ol
I " R T O R T
R
T
Formulation
Figure 3-Intraindividual comparison of the primary pharmacokinetic characteristics of cyclosporine following replicate oral administrations of 300-mg reference (R) formulation and 180-mg test (T) formulation to 24 healthy volunteers.
The doses for each formulation were preselected to provide the same systemic availability of cyclosporine; this was confirmed by comparison of the AUCbs that were comparable (p = 0.57; statistical power 100%) for the test (3424 f 627 h.ng/mL) and the reference (3354 f 816 ha ng/mL) formulations. The terminal disposition half-life was statistically different on the two assessment occasions for the reference formulation (p < 0.01). A between-formulation comparison could not be performed inasmuch as the reference data could not be pooled. Intersubject Variability-Synoptic views of the cyclosporine whole blood concentration-time profiles for the first of the replicate administrations of each formulation are illustrated graphically in Figures 1 and 2. As shown in Table 2, intersubject variability as assessed by ANOVA MSb was significantly less for the pharmacokinetic parameters of the test formulation with the exception of AUCb; for the latter parameter, a trend (p < 0.2) toward lower variability was evident for the test formulation. Intersubject CVs for the primary parameters of cyclosporine ranged from 3 to 22% for the test formulation compared with 20 to 34% for the reference formulation. Intrasubject Variability-Intrasubject variability in the rate and extent of absorption is represented in Figure 3. As compiled in Table 2, comparison of t h e ANOVA MS, demonstrated that the variability within subjects was significantly less for the pharmacokinetic parameters of the test
446 /Journal of Pharmaceutical Sciences Vol. 83, No. 3, M r c h 1994
Sandimmune Neoral is a microemulsion formulation of cyclosporine that represents an improvement in the delivery of the drug for oral absorption. Altering the oral formulation had a demonstrable impact on pharmacokinetic variability. In comparison with the currently marketed formulation, reduced inter- and intraindividual variability in cyclosporine pharmacokinetics from Sandimmune Neoral yields a more consistent and predictable concentration-time profile. It is anticipated that this will translate clinically into greater facility in reaching the desired target blood Concentration range, fewer maintenance dosage adjustments, and perhaps a reduction in the frequency of concentration monitoring necessary in routine patient management.
References and Notes 1. Ptachinski, R. J.; Venkataramanan, R.; Burckart, G. J. Clin. Pharmacokinet. 1986,11, 107-132. 2. Grevel, J. Transplant R o c . 1986, ld(supp1. 5 ) , 9-15. 3. Drewe, J.; Beglinger, C.; Kissel, T. Br. J. Clin. Pharmacol. 1992, 33, 39-43. 4. Reymond, J. P.;Steimer,J. L.; Niederberger,W. J.Pharmacokinet. Biopharm. 1989,16, 331-353. 5. Reymond, J.; Sucker, H. Pharm. Res. 1988,5,673-676. 6. Reymond, *J.; Sucker, H.; Vonderscher, J. Pharm. Res. 1988, 5, 677-679. 7. Trull, A. K.; Tan, K. K. C.; Uttridge, J.; Bauer, T.; Alexander, G. J. M.; Jamieson, N. V. Lancet 1993,341,433. 8. Albert, K. S.; Smith, R. B. In Drug Absorption and Disposition: Statistical Considerations; Albert, K. S., Ed.; American Pharmaceutical Association: Washington, D.C., 1980; pp 87-113. 9. Ball, P. E.; Munzer, H.; Keller, H. P.; Abisch, E.; Rosenthaler, J. Clin. Chem. 1988,34,257-260. 10. Steinijans, V. W.; Hauschke, D. Znt. J. Clin. Pharmacol. Ther. Toxicol. 1990,28, 105-110. 11. RS/1 for VAX and Micro VAX. Release 4.1 software:Bolt Beranek
and Newman Inc. (BBN) Software Products Corporation: Cambridge, MA, 1989.