Population Pharmacokinetics of Cyclosporine in

Chinese Cardiac Transplant Recipients

Ophelia Q. P. Yin, Ph.D., So-Kei Lau, B.S., and Moses S. S. Chow, Pharm.D.

Study Objective. To characterize the population pharmacokinetics of

cyclosporine in Chinese patients undergoing cardiac transplantation and to
identify the demographic and clinical covariates affecting cyclosporine
clearance.
Design. Population pharmacokinetic analysis using data from a retrospective
chart review.
Setting. Specialty hospital in Hong Kong for treatment of cardiac and
pulmonary diseases.
Patients. Thirty-eight Chinese adult patients (mean age 46 yrs) who had
undergone routine cyclosporine therapeutic drug monitoring after cardiac
transplantation between January 1, 1991, and December 31, 2003.
Measurements and Main Results. Data regarding dosing, demographics,
clinical laboratory values, and concurrent drugs were collected
retrospectively. Data were included if patients had blood cyclosporine
concentrations determined for at least 12 weeks after transplantation; an
average of 18 blood samples/patient were collected. Population modeling
was performed using a one-compartment linear model with first-order
absorption and elimination. Various demographic and clinical covariates
were tested for their significant effects on the apparent oral clearance (Cl/F)
of cyclosporine. The stability of the final population model was evaluated
by using the bootstrap resampling method. Statistically significant
associations were observed between Cl/F and each of the following
covariates: body weight (BW), use of diltiazem (DIL), and hematocrit
value (HCT). The final model was Cl/F = 5.00 • (1 – DIL) + 365/HCT +
(0.144 • BW). The interindividual variabilities of Cl/F and apparent
volume of distribution were 14.5% and 40.2%, respectively. The mean
parameter estimates obtained from bootstrap analyses were highly
consistent with those obtained with the original data set.
Conclusion. The estimated Cl/F values of cyclosporine in our Chinese
cardiac transplant recipients appeared to be similar to those reported for
Caucasian cardiac transplant recipients. Thus, our data provide support
that a cyclosporine dosage regimen similar to that in Caucasian patients
may be needed in Chinese cardiac transplant recipients. However, further
studies are required to determine the optimum cyclosporine dosage
regimen in the Chinese population.
Key Words. cyclosporine, drug monitoring, NONMEM software, population
pharmacokinetics, cardiac transplantation.
(Pharmacotherapy 2006;26(6):790–797)
 CYCLOSPORINE KINETICS IN CHINESE CARDIAC TRANSPLANT RECIPIENTS Yin et al
Cyclosporine, a potent immunosuppressive
agent, is widely used for the prevention of graft
rejection in organ transplantation. Cyclosporine
exposure is correlated to clinical outcome after
transplantation. 1 This drug has a narrow
therapeutic index: suboptimal blood cyclosporine
concentration can lead to increased risk of acute
rejection or graft loss, whereas supratherapeutic
concentrations can cause adverse effects such as
nephrotoxicity, hypertension, and diabetes
mellitus.1, 2 Thus, therapeutic drug monitoring to
achieve appropriate cyclosporine concentrations
is needed to ensure the safety and efficacy of the
agent.
Wide inter- and intraindividual variability exist
in cyclosporine pharmacokinetics. A 7–12-fold
interindividual variability in its absorption and
clearance has been reported. 3 Such variability
may be associated with various factors such as
disease, physiologic changes, and concurrent
drug therapy. In the clinical setting, the usual
practice of cyclosporine dosage adjustment is to
determine its concentration frequently and then
adjust the dosage by using a trial-and-error
approach. Such an approach is time consuming
and not optimal in achieving therapeutic
concentrations early after transplantation.4 A
dosing method that can take into account the
various factors that influence cyclosporine
pharmacokinetics is desirable.
Population pharmacokinetic analysis offers an
opportunity to evaluate the pharmacokinetics of
a given drug, by using sparse drug concentration
data to determine the influence of various demo-
graphic and clinical factors (i.e., covariates). 5
Such information from population pharmaco-
kinetic analysis could be helpful to generate an
appropriate cyclosporine dosage regimen to
achieve desirable blood concentrations in
individual patients under different physiologic
and therapeutic conditions.
Although population pharmacokinetics of
cyclosporine have been studied in Caucasian
patients,6–10 little information is available on the
pharmacokinetics of cyclosporine in the Chinese
population, especially cardiac transplant
recipients. Recently, interethnic differences in
From the School of Pharmacy, Faculty of Medicine, The
Chinese University of Hong Kong, Shatin, New Territories,
Hong Kong (Drs. Yin and Chow); and the Department of
Pharmacy, Grantham Hospital, Aberdeen, Hong Kong (Ms.
Lau).
Address reprint requests to Ophelia Q. P. Yin, Ph.D.,
School of Pharmacy, Faculty of Medicine, The Chinese
University of Hong Kong, Shatin, New Territories, Hong
Kong; e-mail: qpyin@cuhk.edu.hk.
791
cyclosporine pharmacokinetics were reported
between African-American and Caucasian
subjects.11 These interethnic differences were
shown to be attributed to the polymorphic
expression of the cytochrome P450 (CYP) 3A4
and 3A5 isoenzymes and P-glycoprotein, which
are known to influence cyclosporine pharmaco-
kinetics. 11 Since the allelic distribution of
CYP3A4, CYP3A5, and P-glycoprotein are
different between Caucasian and Chinese
populations, 12, 13 we postulated that the
pharmacokinetics of cyclosporine might be
different between these populations. Thus, we
conducted this study to characterize the
population pharmacokinetics of cyclosporine in
Chinese cardiac transplant recipients and,
concurrently, to identify the demographic and
clinical covariates affecting cyclosporine
clearance.
Methods
Patient Data
Data were collected from a retrospective review
of the medical charts of Chinese adult patients
who had undergone routine cyclosporine
therapeutic drug monitoring after cardiac
transplantation between January 1, 1991, and
December 31, 2003, at Grantham Hospital, a
specialty hospital for treatment of cardiac and
pulmonary diseases in Hong Kong. Patients were
excluded if they had clinically significant
laboratory abnormalities (e.g., abnormal liver and
renal function levels) or acute rejection, or if they
were receiving other drugs known to interact
with cyclosporine (i.e., that would affect the
absorption and metabolism of cyclosporine).
Patients were enrolled if they received an oral
dose of the microemulsion formulation of
cyclosporine 25–250 mg twice/day (Neoral;
Novartis Pharmaceuticals, East Hanover, NJ) in
combination with azathioprine and prednisone
and had their predose and trough blood
cyclosporine concentrations determined for at
least 12 weeks after transplantation.
For all enrolled patients, cyclosporine blood
samples were collected in the morning before the
next dose. The following data were collected for
each patient during the 12 weeks: date, time,
and postoperative days on which cyclosporine
was administered and on which cyclosporine
blood concentration was obtained; age, sex, and
body weight; hematocrit and serum albumin,
serum creatinine, serum alkaline phosphatase,
aspartate aminotransferase, and alanine
792 PHARMACOTHERAPY Volume 26, Number 6, 2006
aminotransferase levels; and concurrent drug
therapy. The accuracy of these data was checked
by reviewing related records (e.g., computer-
generated laboratory reports, drug dispensing
history, progress notes, and discharge notes).
Blood Sample Analysis
Whole blood cyclosporine concentrations were
determined by a fluorescence polarization
immunoassay with the AxSYM system (Abbott
Laboratories, Abbott Park, IL). The samples
were analyzed according to the Principles of
Good Laboratory Practice. Before each measure-
ment, a standard calibration curve was
constructed over a concentration range of 0–800
ng/ml. The intra- and interday precision
(coefficient of variation) was less than 10%, and
the accuracy ranged from 90–102% for quality-
control samples at concentrations of 70, 300, and
600 ng/ml. The lower limit of quantification was
9.8 ng/ml.
Population Pharmacokinetic Analysis
Population pharmacokinetic analysis was
performed by using nonlinear mixed effects
modeling with the NONMEM program, version
V, level 1.1 (GloboMax LLC, Hanover, MD). The
double precision and first-order conditional
estimation methods were used.
Structural Model Development
A one-compartment linear model with first-
order absorption and elimination was used for
the base model building. The general equation
for calculation of cyclosporine concentration was
as follows:
F • D • Ka
Cp = —————— [e-ket – e-kat]
V • (Ka – Ke)
where D is the dose, V/F is the apparent volume
of distribution, Ka and Ke refer to the absorption
rate constant and elimination rate constant,
respectively, and t refers to time.
Since the data consisted of only trough
cyclosporine concentrations, certain pharmaco-
kinetic parameters were assumed for the
estimation of cyclosporine clearance values.
First, we fixed Ka to be 2.0 hours-1 based on the
literature value reported previously for Chinese
transplant recipients who received cyclosporine
microemulsion formulation.14, 15 Our sensitivity
analysis with different K a values tested from
0.5–5.0 also found that the lowest estimate errors
were obtained by fixing K a to 2.0 hours -1 .
Second, the value of V/F was set as TVV = 4.5 •
BW, where TVV is the population mean value of
V/F and BW is body weight. The appropriateness
of this value was also assessed by a sensitivity
analysis.
Finally, the interindividual variabilities of V/F
and apparent oral clearance (Cl/F) were
estimated by using the exponential error model
(based on an initial comparison among different
models such as additive, proportional,
exponential, or mixed model) according to the
following equations:
V/Fi = TVV • exp(hv,i)
Cl/Fi = TVCl • exp(hCl,i)
where the index i denotes individual and
represents the V/F and Cl/F in the “ith”
individual, and TVV and TVCl are the population
mean values of V/F and Cl/F, respectively. The hi
is the normally distributed random variable, with
mean zero and variance v2.
Various variance models used in NONMEN
(i.e., additive, proportional, or mixed) were also
tested for modeling the intraindividual variability
(random error). A proportional error model was
found to be appropriate for our analysis:
~
Cij = C ij • (1 + eij)
~
where C ij is the jth observed plasma
concentration in the ith individual, C ij is the
corresponding predicted concentration, and eij is
the residual variability term representing random
error with mean zero and variance s2.
Covariate Model Development
The effects of each potential covariate on Cl/F
were systematically examined. These included
age, body weight, sex, coadministration of
diltiazem, postoperative day, hematocrit, and
serum albumin, serum creatinine, alkaline
phosphatase, aspartate aminotransferase, and
alanine aminotransferase levels. The assessment
and selection of significant covariates were based
on the analyses from several steps (see also
Results). First, diagnostic plots and univariate
analysis were performed to initially screen the
potential individual covariates. Next, a forward
adding-on technique was used to evaluate the
relative importance of individual covariates
selected in the first step. Finally, a deletion
process was used to confirm the final covariate
model.
For the above analyses, a decrease in the
objective function value of at least 8 units
 CYCLOSPORINE KINETICS IN CHINESE CARDIAC TRANSPLANT RECIPIENTS Yin et al 793
Table 1. Hematocrit Values in the 38 Chinese Cardiac Table 2. Demographics and Laboratory Data in the 38
Transplant Recipients Patients
Weeks After Hematocrit (%) Characteristic Value
Transplantation Mean ± SD Range 95% CI No. of Patients
0–2 30.0 ± 2.9 24.8–36.5 29.0–30.9 M/F 30/8
2–4 34.0 ± 3.7 25.8–40.6 32.8–35.2 Concomitant diltiazem
4–6 35.1 ± 4.6 26.3–45.2 33.6–36.6 Always 16
6–8 36.6 ± 4.6 28.2–46.4 35.1–38.1 Never 20
8–10 37.4 ± 4.6 29.7–46.7 35.8–38.9 Sometimes 2
10–12 37.8 ± 4.7 30.1–47.9 36.2–39.4 Mean (range)
CI = confidence interval.
Age (yrs) 46 (19–65)
Weight (kg) 59 (43–89)
Cyclosporine data
Postoperative starting day 5 (2–10)
(corresponds to p<0.001 in the log-likelihood- Dose (mg/12 hrs) 117 (25–250)
ratio test) was considered statistically significant, Blood concentration (ng/ml) 260 (10–634)
and thus inclusion of such a covariate in the No. of samples/patienta 18 (10–31)
model was subsequently considered. For each Laboratory data
Hematocrit (%) 35 (25–48)
significant covariate, various submodels (i.e., Serum albumin (g/L) 30 (25–36)
linear and exponential covariate models) were Serum creatinine (mg/dl) 1.07 (0.37–2.23)
also tested, and the covariate model associated Alkaline phosphatase (U/L) 128 (65–567)
with the lowest objective function value was Aspartate aminotransferase (U/L) 38 (13–137)
selected. Alanine aminotransferase (U/L) 85 (26–500)
a
Total number of samples obtained for blood cyclosporine
Since a significant increase in hematocrit value concentrations was 694.
was observed within individual patients over the
12-week period (Table 1), the effect of
hematocrit value on cyclosporine Cl/F was tested
and modeled as a step function as follows: sampling procedure is repeated until the
Cl/Fi = f (HCTi) bootstrap samples generated also consist of N
where HCTi refers to the mean hematocrit value vectors such that Y = (X1*, X 2*, ...X i*, X N*),
at each time interval (i.e., 0–2, 2–4, 4–8, 8–12 where Xi* represents all the observations for the
wks), and f (HCTi) is the function of HCTi. ith randomly selected subjects. In our study, 200
Goodness-of-fit diagnosis was based on the bootstrap samples were generated, and the
changes in objective function value, precision of population pharmacokinetic parameters were
parameter estimates, and visual inspection of the estimated for each of the 200 samples by using
residual plots as well as the plots of predicted the final model. The mean and standard error of
versus observed concentrations. parameter estimates from the bootstrap analyses
were then compared with those from the original
Model Validation data set.
The stability and robustness of the final model Results
were tested by using the bootstrap method (since
a relatively small number of subjects were Patient Characteristics
included in this study, the data-splitting method Demographic data of the 38 patients are
was not used for model testing).16, 17 The general summarized in Table 2. Eighteen patients
principle of the bootstrap method is to repeatedly received diltiazem concurrently, with a mean
sample with replacement from the original data daily dose of 157 mg (range 90–360 mg/day). A
set to produce another data set with the same size total of 694 blood cyclosporine concentrations
but a different combination of subjects. For were determined, with a mean of 18 concen-
example, if the original data set is represented by tration samples/patient. The mean concentration
the set of vectors, Y = (X1, X2,...Xi, XN), where the was 260 ng/ml (range 10–634 ng/ml).
vector Xi represents all the data of the ith subject,
a bootstrap sample is generated by repeated Model Building
random sampling and placement of an N-size
“pseudosample” from the original data set. For Compared with various error models, an
each sampling step, each vector (i.e., X1, X2,...Xi, exponential error model for the interindividual
XN) has an equal probability to be selected. This variability of both V/F and Cl/F and a
794 PHARMACOTHERAPY Volume 26, Number 6, 2006
Table 3. Forward Adding-on and Backward Deletion Processes for Development of the Final Model
Estimate (standard error)
Model Structure OFV DOFV U1 U2 U3 U4
TVCl = U1 7213.1 21.6 (0.96) — — —
TVCl = U1 + U2 • (1 – DIL) 7089.9 -123.2 18.0 (0.91) 5.96 (1.21) — —
TVCl = U1 + U2 • (1 – DIL) + U3/HCT 7016.5 -196.6 5.93 (2.54) 5.15 (1.25) 426 (85.6) —
TVCl = U1 + U2 • (1 – DIL) + U3/HCT +
(U4 • BW) 6977.6 -235.5 0.00 5.00 (1.21) 365 (62.0) 0.144 (0.031)
TVCl = U1 + U3/HCT + (U4 • BW) 7069.0 91.4 0.00 — 437 (60.4) 0.158 (0.033)
TVCl = U1 + U4 • BW 7174.5 196.9 9.61 (3.85) — — 0.217 (0.067)
OFV = objective function value; DOFV = change in OFV; TVCl = population mean value of oral clearance of cyclosporine; DIL = use of
diltiazem (equals 1 with concurrent use, 0 without concurrent use); HCT = hematocrit (%); BW = body weight (in kg).

proportional error model for the residual Univariate analysis showed there were
variability in blood cyclosporine concentrations statistically significant associations between Cl/F
resulted in the smallest objective function value. and each of the following covariates: body
Thus, they were chosen as the base model and weight, use of diltiazem, and hematocrit value.
used in all subsequent analyses. Figure 1A shows The relative importance of the three covariates on
the plot of observed versus base model–predicted Cl/F was further evaluated by both forward
cyclosporine concentrations in study patients. adding-on and backward deletion processes, as
shown in Table 3. The results indicated that each
of the three covariates remained statistically
significant (at p<0.001 level) in both adding-on
A and deletion analyses. Thus, the three covariates
800 were incorporated into the final model, which
resulted in better fitting to the observed
Predicted Concentration (ng/ml)

600
cyclosporine concentrations than the base model
(Figure 1B vs 1A). The residual error generally
fell within 2 units for the final model (Figure 2).
400
The pharmacokinetic estimates for cyclosporine
from the final model are summarized in Table 4.
200 For our patient population, the general
expressions of Cl/F and V/F are as follows:
0 Cl/F = [5.00 • (1 – DIL) + 365/HCT
0 200 400 600 800 + (0.144 • BW)] • exp(0.021)
Observed Concentration (ng/ml)
V/F = 4.5 • BW • exp(0.162)
where DIL is use of diltiazem (which equaled 1
B
800
Predicted Concentration (ng/ml)

5
600 4
3
Weighted Residual

2
400
1
0
-1
200
-2
-3
0 -4
0 200 400 600 800 -5
Observed Concentration (ng/ml) 0 200 400 600

Figure 1. Plots of observed versus predicted cyclosporine Predicted Concentration (ng/ml)

concentrations by the base model (A) and by the final Figure 2. Plot of weighted residuals versus cyclosporine
model (B). Line indicates the line of identity. concentrations predicted by the final model.
 CYCLOSPORINE KINETICS IN CHINESE CARDIAC TRANSPLANT RECIPIENTS Yin et al
Table 4. Population Pharmacokinetic Estimates for Cyclosporine from the Final Model and Bootstrap Validation
Final Model
Parameter Estimate
Cl/F = U1 • (1 - DIL) + U2/HCT + (U3 • BW)
U1 5.00
U2 365
U3 0.144
Interindividual variability in Cl/F (CV%) 14.5
Interindividual variability in V/F (CV%) 40.2
Residual variability (CV%) 30.3
Cl/F = oral clearance (where F = bioavailability); DIL = use of diltiazem (equals 1 with concurrent use, 0 without concurrent use);
HCT = hematocrit (%); BW = body weight (in kg); CV% = coefficient of variation; V/F = apparent volume of distribution.
a
Data are the mean of 200 bootstrap analyses.
with concurrent diltiazem use and zero without
concurrent diltiazem use), HCT is hematocrit
value (%), and BW is body weight (in kg).
The interindividual variabilities (coefficients of
variation) of Cl/F and V/F decreased from 32.6%
to 14.5% and from 50.6% to 40.2%, respectively,
when the three covariates (body weight, use of
diltiazem, hematocrit value) were incorporated
into the final model. The intraindividual
(residual) variability in cyclosporine concentrations
was estimated to be 35.2% with the base model
and 30.3% with the final model. All structural
model parameters as well as variance parameters
were estimated with adequate precision, with
coefficients of variation ranging from 10.0–24.9%
(Table 4).
Model Validation
The mean and standard error of each parameter
estimate obtained from the 200 bootstrap
samples are also summarized in Table 4. These
mean parameter estimates were highly consistent
with those obtained with the original data set.
Discussion
Our population pharmacokinetic analysis
demonstrated that cyclosporine Cl/F in the
cardiac transplant recipient was significantly
related to the use of diltiazem, hematocrit value,
and body weight. Incorporation of these
covariates into the final population model
resulted in significant decreases in the objective
function value (from 7213.1 to 6977.6) as well as
the interindividual variability of cyclosporine
Cl/F (coefficient of variation from 32.6% to
14.5%).
Our model specifically showed a significant
decrease (22.0%) in cyclosporine Cl/F with the
use of diltiazem (Cl/F was estimated to be 0.41
795
Bootstrapa
Standard Error Estimate Standard Error
1.21 5.04 1.17
62.0 362 65.4
0.031 0.145 0.031
9.98 15.2 10.5
24.9 41.1 26.3
11.3 30.0 11.4
L/hr/kg for patients not receiving diltiazem and
0.32 L/hr/kg for those receiving diltiazem). This
observed change is generally consistent with
those changes reported previously in the
literature.18–20 Both diltiazem and cyclosporine
are known to be substrates of CYP3A and P-
glycoprotein, 21–24 and diltiazem also has been
demonstrated to be an inhibitor of CYP3A and P-
glycoprotein.25 Thus, the decrease in Cl/F could
reflect a decrease in liver and/or gut metabolism,
as well as an increase in absorption of
cyclosporine.25
Another specific significant effect is related to
hematocrit value. Our analysis showed that
hematocrit value also contributed significantly to
the interindividual variability of cyclosporine
Cl/F (objective function value decreased by 100.3
units and interindividual variability decreased by
5.1% after addition of hematocrit value as a
covariate). In addition, the intraindividual
variability in hematocrit value (change with
time) appeared to be associated with the change
in cyclosporine Cl/F over time. In our patient
population, cyclosporine Cl/F was estimated to
be 0.44 L/hour/kg at week 1 but 0.39 L/hour/kg
at week 12, with a change in hematocrit value
from 30.0% to 37.8% after transplantation.
Proportionally, a 26.0% increase in hematocrit
value over a 12-week period was found to be
related to an 11.4% decrease in cyclosporine Cl/F.
Our finding on the relationship between
hematocrit value and cyclosporine Cl/F is
consistent with results from other studies, in
which patients with high hematocrit values were
found to have high cyclosporine exposure as well
as low cyclosporine Cl/F.8, 26, 27 However, results
of another study 28 showed a lack of such
relationship. The reasons for the conflicting data
are unclear. Nevertheless, cyclosporine is known
to be concentrated in the blood cells. With an
796 PHARMACOTHERAPY Volume 26, Number 6, 2006
increase in hematocrit value, more cyclosporine
will be bound to erythrocytes, leading to an
increase in the observed total blood concen-
tration of cyclosporine. Both our inter- and
intraindividual data support this finding.
The optimum dosage of cyclosporine in
Chinese cardiac transplant recipients in
comparison to other ethnic populations is
unknown; however, our study provided some
preliminary information. In Caucasian patients
undergoing cardiac transplantation, the average
cyclosporine Cl/F ranges from 0.34–0.44
L/hour/kg, based on the rich data obtained from
pharmacokinetic studies of the cyclosporine
microemulsion formulation.15, 29 The typical Cl/F
value in our Chinese patients was 0.41
L/hour/kg, which is consistent with the above-
reported values. Our data appear to support that
a cyclosporine dosage regimen similar to that for
Caucasian patients undergoing cardiac transplan-
tation would be needed for Chinese cardiac
transplant recipients. However, further studies
are required to confirm this finding.
A limitation of this study is that our data
consisted of only trough blood cyclosporine
concentrations, and we used a one-compartment
model for describing cyclosporine pharmaco-
kinetics. Although the use of more complex
pharmacokinetic models (e.g., two-compartment
with or without first-pass model, g-distribution
absorption model) appears to be superior for
cyclosporine,30–32 the one-compartment model
has been adopted in a number of studies and
appears also to be satisfactory.9, 10, 33
Another limitation is the relatively small
number of subjects included in our study and the
lack of sufficient subjects to allow analysis of
other potential covariates that may influence
blood cyclosporine concentrations. These
include cholesterol concentration and the use of
other concurrent drugs (e.g., antifungal agents,
macrolide antibiotics, anticonvulsants). Since
the number of patients receiving these drugs was
very small in our population, these covariates
might be less important from a population
perspective. However, individual patients
receiving these drugs concurrently with
cyclosporine should be monitored carefully.
Conclusion
With our population pharmacokinetic analysis,
we were able to describe observed cyclosporine
concentrations and identify significant
demographic and clinical factors contributing to
the variability in cyclosporine clearance in
Chinese cardiac transplant recipients. The
estimated Cl/F values of cyclosporine in our
Chinese patients appear to be similar to those
reported for Caucasian cardiac transplant
recipients. Thus, our data provide support that a
cyclosporine dosage regimen similar to that in
Caucasian patients may be needed in Chinese
cardiac transplant recipients. However, further
studies are required to determine the optimum
cyclosporine dosage regimen in the Chinese
population.
References
1. Shaw LM, Bowers L, Demers L, et al. Critical issues in
cyclosporine monitoring: report of the task force on
cyclosporine monitoring. Clin Chem 1987;33:1269–88.
2. Hesselink DA, Smak Gregoor PJ, Weimar W. The use of
cyclosporine in renal transplantation. Transplant Proc
2004;36:S99–106.
3. Lindholm A, Kaham BD. Influence of cyclosporine pharmaco-
kinetics, trough concentrations, and AUC monitoring on
outcome after kidney transplantation. Clin Pharmacol Ther
1993;54:205–18.
4. Clase CM, Mahalati K, Kiberd BA, et al. Adequate early
cyclosporin exposure is critical to prevent renal allograft
rejection: patients monitored by absorption profiling. Am J
transplant 2002;2:789–95.
5. Aarons L, Balant LP, Mentre F, et al. Population approaches in
drug development: report on expert meeting to discuss
population pharmacokinetic/pharmacodynamic software. Eur J
Clin Pharmacol 1994;46:389–91.
6. Anderson JE, Munday AS, Kelman AW, et al. Evaluation of a
Bayesian approach to the pharmacokinetic interpretation of
cyclosporine concentrations in renal allograft recipients. Ther
Drug Monit 1994;16:160–5.
7. Ptachcinski RJ, Venkataramanan R, Rosenthal JT, et al.
Cyclosporine kinetics in renal transplantation. Clin Pharmacol
Ther 1985;38:296–300.
8. Venkataramanan R, Burchart GJ, Ptachcinski RJ, et al.
Cyclosporine pharmacokinetics in heart transplant patients.
Transplant Proc 1986;18:768–70.
9. Tan Trull KKCAK, Hue KL, Best NG, Wallwork J,
Higenbottam TW. Pharmacokinetics of cyclosporine in heart
and lung transplant candidates and recipients with cystic
fibrosis and Eisenmenger’s syndrome. Clin Pharmacol Ther
1993;53:544–54.
10. Parke J, Charles, BG. NONMEM population pharmacokinetic
modeling of orally administered cyclosporine from routine drug
monitoring data after heart transplantation. Ther Drug Monit
1998;20:284–93.
11. Dirks NL, Huth B, Yates CR, Meibohm B. Pharmacokinetics of
immunosuppressants: a perspective on ethnic differences. Int J
Clin Pharmacol Ther 2004;42:701–18.
12. Ieiri I, Takane H, Otsubo K. The MDR1 (ABCB1) gene poly-
morphism and its clinical implications. Clin Pharmacokinet
2004;43:553–76.
13. Hustert E, Haberl M, Burk O, et al. The genetic determinants
of the CYP3A5 polymorphism. Pharmacogenetics 2001;11:
773–9.
14. Akhlaghi F, Keogh AM, McLachlan AJ, Kann A.
Pharmacokinetics of cyclosporine in heart recipients receiving
metabolic inhibitors. J Heart Lung Transplant 2001;20:431–8.
15. Balram C, Sivathasan C, Cheung YB, Tan SB, Tan YS. A
limited sampling strategy for the estimation of 12-hour Neoral
systemic drug exposure in heart transplant recipients. J Heart
Lung Transplant 2002;21:1016–21.
16. Ette EI. Stability and performance of a population
 CYCLOSPORINE KINETICS IN CHINESE CARDIAC TRANSPLANT RECIPIENTS Yin et al
pharmacokinetic model. J Clin Pharmacol 1997;37:486–95.
17. Parke J, Holford NHG, Charles BG. A procedure for generating
samples for the validation of nonlinear mixed-effects 26.
population models. Comput Methods Programs Biomed
1999;59:19–29.
18. Masri MA, Shakuntala V, Shanwaz M, et al. Pharmacokinetics
of cyclosporine in renal transplant patients on diltiazem. 27.
Transplant Proc 1994;26:1921.
19. Jones TE, Morris RG, Mathew TH. Diltiazem-cyclosporin
pharmacokinetic interaction: dose-response relationship. Br J
Clin Pharmacol 1997;44:499–504. 28.
20. Mclachlan AJ, Tett SE. Effect of metabolic inhibitors on
cyclosporin pharmacokinetics using a population approach.
Ther Drug Monit 1998;20:390–5. 29.
21. Pichard L, Gillet G, Fabre I, et al. Identification of the rabbit
and human cytochromes P-450IIIA as the major enzymes
involved in the N-demethylation of diltiazem. Drug Metab
Dispos 1990;18:711–19.
22. Molden E, Christensen H, Sund RB. Extensive metabolism of 30.
diltiazem and P-glycoprotein-mediated efflux of desacetyl-
diltiazem by rat jejunum in vitro. Drug Metab Dispos
2000;28:107–9.
23. Kronbach T, Fiseher V, Meyer UA. Cyclosporine metabolism in 31.
human liver: identification of a cytochrome P-450III gene
family as the major cyclosporine metabolizing enzyme
explaining interactions of cyclosporine with other drugs. Clin
Pharmacol Ther 1988;43:630–5. 32.
24. Lown KS, Mayo RR, Leichtmann AB, et al. Role of intestinal p-
glycoprotein (mdr1) in interpatient variation in the oral
bioavailability of cyclosporine. Clin Pharmacol Ther 33.
1997;62:248–60.
25. Watkins PB, Leichtman A. The molecular basis of cyclosporine
797
A metabolism, pharmacokinetics and drug interactions. Graft
1999;2:177–81.
Lindholm A, Welsh M, Alton C, Kahan BD. Demographic
factors influencing cyclosporine pharmacokinetic parameters in
patients with uremia: racial differences in bioavailability. Clin
Pharmacol Ther 1992;52:359–71.
Yee GC, McGuire TR, Gmur DJ, Lennon TP, Deeg HJ. Blood
cyclosporine pharmacokinetics in patients undergoing marrow
transplantation: influence of age, obesity and hematocrit.
Transplantation 1988;46:399–402.
Grevel J, Reynolds KL, Rutzky LP, Kahan BD. Influence of
demographic factors on cyclosporine pharmacokinetics in adult
uremic patients. J Clin Pharmacol 1989;29:261–6.
Hesselink DA, van Gelder T, van Schaik RH, et al. Population
pharmacokinetics of cyclosporine in kidney and heart
transplant recipients and the influence of ethnicity and genetic
polymorphisms in the MDR-1, CYP3A4, and CYP3A5 genes.
Clin Pharmacol Ther 2004;76:545–56.
Schadeli F, Marti HP, Frey FJ, Uehlinger DE. Population
pharmacokinetic model to predict steady-state exposure to
once-daily cyclosporine microemulsion in renal transplant
recipients. Clin Pharmacokinet 2002;41:59–69.
Ducharme MP, Verret L, Brouillette D, Sirois G. Ability of a
first-pass pharmacokinetic model to characterize cyclosporine
blood concentrations after administrations of Sandimmune or
Neoral formulations. Ther Drug Monit 1998;20:165–71.
Debord J, Risco E, Harel M, et al. Application of a Gamma
model of absorption to oral cyclosporine. Clin Pharmacokinet
2001;40:375–82.
Rosenbaum SE, Baheti G, Trull AK, Akhlaghi F. Population
pharmacokinetics of cyclosporine in cardiopulmonary
transplant recipients. Ther Drug Monit 2005;27:116–22.