epilepsy , pharmacokinetics , study , design , bipharmaceutics , health , medicine , pharmacy , clinical pharmacy , pakistan , lahore , dose dependent clearance , population pk

1. Population pharmacokinetics of
phenytoin in japanese patients with
epilepsy:Analysis with a dose
dependent clearence model
Sidra Ali
Saima tufail
IPS,UVAS Lahore

2. Epilepsy
• Epilepsy is not a single entity .
• Instead its an assortment of different seizure
types and syndromes originating from several
mechanism that have in common the sudden
,excessive ,and synchronous discharge of
cerebral neurons.
• The abnormal electrical activity results in a
variety of events,including

3. • Loss of consciousness
• Abnormal movements
• Atypical or odd behaviour
• Distorted perceptions that are of limited
duration but recur if untreated.

4. Phenytoin
• It blocks the voltage gated sodium channels by
selectively binding to the channel in the
inactive state and slowing its rate of recovery.
• At very high concentration,it can also block
voltage dependent calcium channels and
interferes with the release of monoaminergic
neurotransmitter.
• Effective for partial seizures and generallized
tonic clonic seizures and status epilepticus.

5. Pharmacokinetics
• 90 % bound to plasma protein
• Metabolized by CYP2C ,CYP3A families and
UGT enzyme system.
• Insoluble weak acid, administered as a sodium
salt.
• Follows non linear (zero order) kinetics.
• Clearence 10L/h
• Vd 70 L

6. • Half life 4 hr
• Distributes into body tissues within 30-60
minutes.
• Metabolites excreted in urine.
• Metabolism altered such as hepatic cirrhosis
or concomitent drug administration.

7. • Small increase in daily dose produce large
increases in plasma concentration, resulting in
drug induced toxicity.
• Depression of CNS occurs particularly in the
cerebellum and vestibular system,causing
nystagmus and ataxia

8. • Side effects
• Gingival hyperplasia may cause the gums to
grow over the teeth
• Long term use may lead to peripheral
neuropathies and osteoporosis.
• Fosphenytoin is a prodrug ,rapidly converted
to phenytoin in the blood ,reaching high levels
within minutes.

9. • May be administered IM
• Phenytoin sodium should never be
administered IM b/C it can cause tissue
damage and necrosis.

10. Model used
• Michaelis menten model
• Lineweaver burk form of MM model
• Dose dependent clearence model
• NONMEM (nonlinear mixed effects model.

11. Michaelis menten model
• MM constant km” shows the concentration of
substrate when the reaction velocity is equall
to one half of the maximal velocity for the
reaction.
• Can also be thought as a measure of how well
a substrate complexes with a given enzyme.
• MM hypothesis assumes the rapid ,reversible
formation of a complex between an enzyme
and its substrate.

12. Lineweaver burk model
• Also called double reciprocal plot is a graphical
representation of lineweaver burk equation of
enzyme kinetics,described by Hans lineweaver
and Dean burk in 1934.
• Widely used to determine important terms in
enzyme kinetics, such as Km and Vmax
• Vmax = max reaction velocity
• Km= michaelis menten constant

13. • The y-intercept of such graph is equivalent to
the inverse of vmax
• X-intercept represents the -1/km
• Used for the representation of kinetic
data,non linear regression or alternative forms
of MM equation.

15. Introduction
• Population analysis of dose dependent pharmacokinetic
parameters of phenytoin have been performed using MM or
Lineweaver-burk form of MM,where dosing rate is assumed
to be equivalent to the elimination rate at steady state after
repetitive oral administration.The MM provide estimates of
elimination rate and MM constant but no estimate of
Vd.Therefore ,the bayesian regression programe is not
applicable to many clinical situations where steady state
concentrations of phenytoin are not available.
• However population pharmacokinetics of phenytoin have not
been estimated by a one compartment model with MM
elimination due to difficulty in population pharmacokinetic
analysis using explicit solutions.

16. True model versus DDCL model
• Simulation study for the
population analysis
showed that TRUE model
requires precise analysis
method for estimation of
pharmacokinetics
parameters &
considerable
computational time to
estimate mean
population parameters
• One compartment model
with DDCL runs 20 fold
faster than the true and
gives accurate population
parameters for the steady
state pharmacokinetic
data having long half life
relative to dosing interval.
• DDCL is a promising
alternative to TRUE for
estimating population
parameters of drug
exhibiting MM kinetics.

17. sampling
• The population pharmacokinetic parameters
of phenytoin were estimated using routine
TDM data from 116 patients.the 531 serum
concentrations after repetitive oral
administration at steady state were analyzed
using NONMEM.
• One compartment model was used for
pharmacokinetics analysis.

18. • Volume of distribution was estimated to be
1.231/kg in a 42kg patient assuming that the
bioavailability of orally administered is 100%.
• Elimination rate constant and Michaelis –
menten constants were 9.80mg/dl/kg and
9.19 micogram/ml respectively.
• Parameter of power function of weight to
adjust V and Vmax was estimated to be 0.463

19. • Elimination rate constant for phenytoin
appeared to be 16% increased in patients
receiving zonisamide concurrently.
• Population parameters will be helpful for
designing dosage regimens in epileptic
patients.

20. Methodology
• Patients and data description
• Pharmacostatistical model

21. Patients and data description
• Serum phenytoin concentration data at steady
state after repetitive dosing(for more than 4
weeks) were collected from 116 epileptic
patients, who were treated with oral dosage
form.
• Table 1 depicts the characteristics of 116
patients.
• Most of the patients were outpatients,and no

22. patient had renal /hepatic failure.the age of
patients range from 1-37 years old.

23. Distribution of age

24. • The serum concentration of phenytoin was
measured by flourescence polarization
immunoassay.531 serum conc were used for
pharmacokinetic analysis ,263 serum samples
were obtained for the measurement of peak
levels at 2.5 to 4 hours after dosing interval
while 54 samples were obtained for the
measurement of trough levels at 8 to 24 hours

25. • No significant absorption phase observed
following oral administration in the serum
phenytoin data.since serum samples were
obtained for routine therapeutic drug
monitoring purposes.

26. Pharmacostatistical Model
• The time course of drug concentration in an interdose
interval at steady state after repetitive dosing can be
analyzed using DDCL model.
• It is approximate solution to TRUE model(one
compartment model with michaelis-menten
elimination)
• Dij=dose, Tij= dosing interval , Vmaxi= max elimination
rate constant, Kmi= michaelis-menten contant
• It is equation for clearance of phenytoin in “ith”
subject. (SPCij)

27. • Assuming bioavailability 100% ,one compartment
model with rapid absorption during oral
administration.
• Vi= volume of distribution , Tij= time after doing
associated with SPCij , εij= random variable
• DDCL model gives exact solution to TRUE for peak
and trough levels at steady state on repetitive
bolus administration.

28. • Influence of body weight on Vmaxi and Vi of phenytoin and the
interindivisual variabilities of Vmaxi and Vi were modeled.
• Modelling body size allows the data to determine the optimal
power function of weight to be used to adjust Vmaxi and Vi.
• SIZEi=hypothetical body size to adjust vmaxi and vi in ith indivisual
• θwt= 1 SIZEi is simply linealry proportional to weight .
• θwt= 0 SIZE I is independent of weight

29. Therapeutic drug monitoring
• Serum phenytoin concentration was high
when coadministered zonisamide.
• Influence of zonisamide on Kmi of phenytoin
and interindivisual variability of Kmi were
modeled.

30. • ZNSi = 1 for co administered zonidamide.
• ZNSi = 0 for no zonidamide administered.
• Θzns = parameter to estimate effect of
zonidamide on phenytoin clearance
• ηkmi= random variable describing
interindivisual variability of kmi with zero
mean and variance equal to ω2km.

31. Data Analysis
• Performed with NONMEM software at koyoto
university data processing center.
• It pools data from all indivisuals but explicity models
and handels the complicated error structure arising
from a proper accounting of the interindivisual (ηi) and
intraindivisual (εij) random effects.
• It allows us to estimate population mean parameters,
Vmax , km ,V , θwt , θzns.
• Interindivisual variabilities includes ω2Vmax , ω2km ,
ω2V.
• And intraindivisual variabilities σ2.

32. • In thi study we used first order conditional
estimation method with η,ε interaction.
• NONMEM provides estimates of the standard
error for all the parameters, and standard error
can be used to define confidence intervals (CI) for
true parameter values.
• Statistical significance of the parameters was also
evaluated by the likelihood ratio test using the
minimum value of the objective function
produced by NONMEM.

33. • If hypothetical value was greater than 3.84,
the parameter value was considered to be
statistically significant.

34. Results

35. Discussion
• The general solution for the population analysis with
the TRUE model has been available with ADVAN10 in
the PREDEPP library of NONMEM software.
• TRUE model require 50 fold longer computational time
than that with the MM model.
• 1st order method Is not adeqaute for the TRUE model
which is highly nonlinear in its parameters.
• The population analysis with the TRUE model requires
precise analysis method(e.g 1st order conditional
estimation method) it demands 15 fold CPU time than
the 1st order method.

36. • Mainly for this reason pop-pk parameters of
phenytoin have not been estimated using TRUE
model.
• On the other hand , the DDCL model was shown
to be a very precise solution for TRUE model.
• It was firstly applied for estimation of pop-pk
parameters of an antiepileptic drug ,zonisamide.
• The ddcl model in conjuction with precise
analysis method gives accurate population
parameters as well as the TRUE model, but runs
much faster than TRUE model.

37. • In the present study we apply DDCL model for
estimation of pop-pk parameters of phenytoin.
• CPU time required for the estimation of final
parameters and their standard error wa approx.
200s on our 276-MIPS mainframe machine.
• The MM model gives the downwardly biased
estimates of Vmax and Km for the TRUE model ,
especially when the pk data are analyzed by the
1st order method.
• The Vmax and Km values of phenytoin estimated
here using DDCL model were significantly larger
than those estimated previously using MM model
or its variants .

38. • Estimated value of V was 1.231/kg in a typical
42kg patient which may be consistent with the
previously reported value 1.01/kg in a typical
70kg patient.
• In present study samples obtained after 2-6 hr.
after dosing and only 54 samples were
obtained for the measurement of trough
levels.

39. • 531 serum concentrations plotted vs. weight corrected daily dose
• Serum phenytoin concentration is lower in younger patient than in older
patient.

40. • Show the relationship between serum phenytoin concentration and size
corrected daily dose, assuming that Vmaxi and Vi were adjusted by using
weight to the power 0.462(θwt not equal to 1)
• The relationship between the serum phenytoin concentration and size
corrected daily dose in the younger patient was almost identical to that in
the older patient.

41. • Effect of coadministered zonisamideon the relationship between the serum
concentration and the size corrected daily dose of phenytoin.
• Slight but significant increase in the serum concentration were observed in patient
receiving zonisamide.

42. • We modeled Km so that the concurrently
administered zonisamide may inhibit the
phenytoin metabolism competitvely.
• Clearance of phenytoin 14% decrease by
zonisamide and serum concentration 16%
increase in patients receiving zonisamide.
• In addition this effect of zonisamide was
considered to be significant (p<0.01)
• The NONMEM analysis revealed a decreased
fit model

43. Conclusion
• Pop-pk of phenytoin was studied using routine TDM
data from 116 epileptic patients.
• The 531 serum concentrations at steady state after
repetitive oral administration were analyzed using a
one compartment model with dose dependent
clearance.(MM elimination)
• The values of V , Vmax , Km , effect of body weight and
concurrently administered zonisamide on the
parameters were estimated simultaneously with
NONMEM program.
• The pop-pk parameter of phenytoin will be useful for
designing dosage regimen in epileptic patient.