DRUG EVALUATION IN PEDIATRICS USING K-PD MODELS: PERSPECTIVES - - PDF document

Tod / Girard 1 EMEA 14 April 2008

DRUG EVALUATION IN PEDIATRICS USING K-PD MODELS: PERSPECTIVES

Michel TOD, PhD Pascal Girard, PhD EA3738, Faculté Médecine Lyon Sud Lyon I University, France

Tod / Girard 2 EMEA 14 April 2008

Describe quantitatively drug kinetics. Simulate and predict. Plan and design clinical trials. Bayesian adaptation of drug dosing.

USES OF MODELS

Tod / Girard 3 EMEA 14 April 2008 Invasiveness:

blood samples for PK.

Logistic

and cost associated with samples and measurements

DRAWBACKS OF PK-PD MODELS

Tod / Girard 4 EMEA 14 April 2008 Kinetic

– Pharmaco-Dynamic model

Drug concentrations are not

measured

Only the kinetics of response is

measured.

A simple model is

used to describe drug concentration kinetics.

A SOLUTION: THE K-PD MODEL

Tod / Girard 5 EMEA 14 April 2008

Body

temperature, heart rate, blood pressure, respiratory peek flow …

Scores for depression (HAMD,…), pain (VAS), … Frequency of seizures, emesis, … ECG, EEG Bone density, tumor size, …

LESS INVASIVE MEASUREMENTS OF THERAPEUTIC RESPONSE

Tod / Girard 6 EMEA 14 April 2008 Simplified PK model:

variable of interest: Input Rate (t) in mg/h IR(t) = Ke.A(t)

COMPONENTS OF A K-PD MODEL (1)

Biophase A(t) Ke Dosing history Time A(t) 5 Log2 / Ke

Tod / Girard 7 EMEA 14 April 2008 Effect model: links IR(t) to E(t)

COMPONENTS OF A K-PD MODEL (2)

) t ( IR EDK ) t ( IR . max E ) t ( E

50 +

= EDK50 = CL.CE50 in mg/h

IR E Emax EDK50

Tod / Girard 8 EMEA 14 April 2008 Model for a continuous response: links E(t) to R(t)

COMPONENTS OF A K-PD MODEL (3)

) t ( R . Kout ) ) t ( IR EDK ) t ( IR . max E 1 .( Kin dt ) t ( dR

50

− + − =

Example: inhibition of production

Tod / Girard 9 EMEA 14 April 2008

TYPICAL CURVES OF 2 K-PD MODELS

K-PD MODEL (INHIBITION OF PRODUCTION) 5 10 15 20 25 30 35 12 24 36 48 60

TIME (H) RESPONSE 25 mg BID 120 mg BID 1200 mg BID

K-PD MODEL (INHIBITION OF LOSS) 20 40 60 80 100 120 12 24 36 48 60

TIME (H) RESPONSE 25 mg BID 120 mg BID 360 mg BID

Tod / Girard 10 EMEA 14 April 2008 Model for categorical response: links E(t) to probability to

• bserve score k of response R(t)

COMPONENTS OF A K-PD MODEL (4)

) t ( IR EDK ) t ( IR . max E B )] k ) t ( R ( P [ it log

50 k

+ ± = ≤

Tod / Girard 11 EMEA 14 April 2008 K-PD MODEL (DIMINUTION OF PROBABILITY)

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 12 24 36 48 60

TIME (H)

• PROB. RESPONSE

P(R = 0) P(R <= 1) P(R <= 2)

TYPICAL CURVES FOR A K-PD MODEL OF CATEGORICAL RESPONSE

Score from 0 to 3

Tod / Girard 12 EMEA 14 April 2008

Typical fits for the K-PD and the PK-PD models in arbitrarily chosen subjects. Observation (o), individual prediction (——), population prediction (----)

K-PD versus PK-PD MODELS

A semimechanistic and mechanistic population PK–PD model for biomarker response to ibandronate, a new bisphosphonate for the treatment of osteoporosis. G. Pillai, 2004 K-PD PK-PD

Tod / Girard 13 EMEA 14 April 2008

Modelling Response Time Profiles in the Absence of Drug Concentrations: Definition and Performance Evaluation

• f the K–PD Model.
• P. Jacqmin et al.
• J. Pharmacokin Pharmacodyn 2007

NEFA plasma concentration–time profiles after IV infusion of N6-(p- sulfophenyl) adenosine in Wistar rats.

K-PD MODEL: continuous response

Tod / Girard 14 EMEA 14 April 2008

Pharmacokinetic/Pharmacodynamic and Time-to-Event Models of Ribavirin- Induced Anaemia in Chronic Hepatitis C

• M. Tod et al.
• Clin. Pharmacokinet. 2005

Prediction of the K-PD model for a typical patient.

K-PD MODEL: continuous response

Tod / Girard 15 EMEA 14 April 2008

K-PD SET-POINT MODEL (1)

Box-plot of MADRS scores of patients treated with clomipramine and placebo

• r lithium. (clinical data)

A kinetic-pharmacodynamic model for clinical trial simulation of antidepressant action: Application to clomipramine–lithium interaction.

• B. Gruwez et al., Contemp Clin Trials, 2007.

Non-invasive measurements …

Tod / Girard 16 EMEA 14 April 2008

K-PD SET-POINT MODEL (2)

MADRS scores of patients treated with clomipramine and placebo or

• lithium. (clinical data)

A kinetic-pharmacodynamic model for clinical trial simulation of antidepressant action: Application to clomipramine–lithium interaction. B.Gruwez et al., Contemp Clin Trials, 2007.

Oscillating profile…

Tod / Girard 17 EMEA 14 April 2008

KPD MODEL: biphasic kinetics

PKPD Modeling of the Effect of Triamcinolone Acetonide on Central Macular Thickness in Patients with Diabetic Macular Edema

• F. Audren et al.,

Invest Ophthalmol Vis Sci. 2004 Examples of individual central macular thickness (CMT) curves calculated from individual CMT values (circles).

Non-invasive measurements …

Tod / Girard 18 EMEA 14 April 2008

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 5 10 15 20 25 30 0.0 0.25 0.50 0.75 1.00

1 2

Week Predicted probability

Observed HFS

5 10 15 20 25 Weekly drug rate (g/week)

Evolution with time of observed HFS scores, weekly drug rate and predicted probabilities of grade 0, 1 and ≥2 predicted by the model in a patient. A predictive model of Hand-and- Foot Syndrome dynamics in patients receiving capecitabine.

• E. Hénin et al, 2007

K-PD MODEL: catégorical data

Tod / Girard 19 EMEA 14 April 2008 The drug PK in biophase is handled as monocompartmental:

• PK is actually monocompartmental, or :
• Effect kinetics is slow compared to drug kinetics (Kout < Ke)
• Complicated response models may be handled if correctly

specified.

K-PD

models for drug-drug interaction are merely identifiable.

LIMITS OF THE K-PD MODEL

Tod / Girard 20 EMEA 14 April 2008 K-PD models have been useful for modelling animal or

human data in adults.

Well suited if effect kinetics is rate limiting Might be used in pediatrics to reduce experimental workload. More

useful if coupled with a minimally invasive measurement of response.

CONCLUSIONS