Approaching the In Silico Child

1
Approaching the In Silico Child Jeffrey S.
Barrett, PhD, FCP
2
Outline

• Background
• Pediatric Pharmacotherapy Defined
• Whats missing?
• Pediatric Priors where do they come from?
• Models for understanding vs prediction
• The EMR -- leveraging hospital informatics
• The Pediatrics Knowledgebase (PKB) Project
• Design Issues
• Methotrexate Drug Dashboard
• Vision for the Future

3
Pharmacotherapy

• Principally concerned with the safe and effective
  management of drug administration.
• Implies an understanding of pharmacokinetics (PK)
  and pharmacodynamics (PD) so that individual
  dosing guidance, when necessary, can be provided
  to optimize patient response within their
  individual therapeutic window.

4
Pharmacotherapy

• 75 prescription drugs in children off-label
• Usage not described in package insert
• Approved indications
• Adequate controlled studies
• Consequences of off label usage
• Benefit, No effect, Harm

5
Pharmacotherapy

• Unapproved is not improper
• Decision based on safety/efficacy data
• Medical literature vs Regulatory Guidance
• Best medical judgment

6
PharmacotherapyThe Landscape for Predicting
Exposure
Urine, Feces, Expired Air
Active/inactive metabolites
ABSORPTION
- Site (i.e., GIT, skin, tissue depot) -
First-pass effect (oral) - Drug properties (i.e.,
solubility)
METABOLISM
ELIMINATION

• Pathway(s)
• Sites (GIT, liver, lung)

- Unchanged drug - Metabolites
Excretory Sites
Distribution in Blood Cells
Bound to plasma proteins
Free Drug in Plasma or Extracellular Fluid
SITE(S) FOR THERAPEUTIC EFFECT(S)
Pharmacologic Activity
DISTRIBUTION
- Sites (Tissues, fat, etc) - Binding
SITE(S) FOR TOXIC EFFECT(S)
Toxic Activity
7
PharmacotherapyWhats Missing?

• Drug disposition in children is best described
  using the term variable
• In general, variability is much greater in first
  3 months of life and declines to adult
  variability
• Estimating exposure is challenging due to
  developmental changes affecting absorption,
  distribution and biotransformation
• Exposure also function of exogenous influences
  (diet, concurrent illness)

J. Steven Leeder, Pharm.D., Ph.D.
8
PharmacotherapyWhats Missing?

• Scaling pediatric from adult dosing data needs
  to take into consideration
• Knowledge of relative contribution of ADME
  components at each developmental stage
• For biotransformation, knowledge of fractional
  contribution of each pathway to total CL
• Isoform-specific patterns of development
• Interindividual variability in the rate and
  pattern of pathway development
• Age-dependent differences in population
  variability

J. Steven Leeder, Pharm.D., Ph.D.
9
Pediatric PriorsAbsorption
J. Steven Leeder, Pharm.D., Ph.D.
10
Pediatric Priors Distribution
Intracellular Water
Protein
Fat
Other
Extracellular Water
Premature
Newborn
4 mos
12 mos
24 mos
36 mos
Adult
20
100
0
40
60
80
Percentage of Total Body Weight
11
Pediatric Priors Metabolism

• Functional drug biotransformation capacity
  acquired in isoform-specific patterns
• Onset in Days CYPs 2C9, 2D6, 2E1 UGTs 1A and
  2B7?
• Onset in Weeks CYP3A4
• Onset in Months CYP1A2
• Onset in Years FMO3

J. Steven Leeder, Pharm.D., Ph.D.
12
Pediatric Priors Metabolism

• Time to activity peaks also isoform-dependent,
  but less well characterized
• In general, in vitro studies indicate that
  variability is much greater in first 3 months of
  life and declines to adult variability
• Newborns at particularly high risk for
  concentration-dependent toxicity due to
  developmentally delayed drug metabolism (e.g.
  chloramphenicol, SSRIs)

J. Steven Leeder, Pharm.D., Ph.D.
13
Pediatric PriorsMetabolism
Liver MassBody Weight Change with Age
Liver Mass ( Body Weight)
Age (years)
J. Steven Leeder, Pharm.D., Ph.D.
14
Pediatric Priors Metabolism
Activity
Toddler
Newborn
Puberty
Adult
J. Steven Leeder, Pharm.D., Ph.D.
15
Pediatric Priors Models for Understanding vs
Prediction
MODEL IMPACT
INFORMATION CONTENT

• Discovery
• Define functional relationships
• PK/PD Data signature
• Early CUI

• Decision-making
• Candidate screening / selection
• Dose selection
• Study designs
• Compound progression

• Patient Pharmacotherapy
• Dosing guidance
• Patient management of AE / ADRs
• Optimize sub-therapeutic response
• Rescue therapy

Discovery
Decision- Making
Pharmacotherapy
16
Pediatric Priors Tools for Prediction
PLASMA FLOW
PLASMA
HEART
HEPATIC ARTERY
SPLEEN
LIVER
BILE
KIDNEY
URINE
BONE MARROW
MUSCLE
CARCASS
17
Pediatric Priors Electronic Medical Records

• Paper-based records have been in existence for
  centuries and their gradual replacement by
  computer-based records has been slowly underway
  for over 20 years.
• The penetration of electronic medical records
  (EMRs) may have reached over 90 in primary care
  practices in Norway, Sweden and Denmark (2003),
  but has been limited to 17 of physician office
  practices in the USA (2001-2003).
• The EMR systems that have been implemented have
  been used primarily for administrative rather
  than clinical purposes.

18
Electronic Medical Records CHOP Environment

• EpicCare and EpicWeb ambulatory computerized
  medical record.
• Sunrise Clinical Manager impatient clinical
  order entry, charting, charging, and
  documentation.
• Wellsoft Emergency Department patient
  management, clinical documentation, and
  reporting.
• ChartMaxx legal medical record for impatient,
  emergency, ambulatory surgery.
• IDX Rad radiology patient management and
  transcription.
• Meditech laboratory information system

19
Pediatric Knowledgebase (PKB)Concept

• A physician-designed informatics system which
  surfaces the most relevant data to guide
  individual patient pharmacotherapy
• Construction of individual drug dashboards
  which provide quantitative prediction (as
  requested) relative to historical and comparative
  patient metrics.

20
Pediatric Knowledgebase (PKB)Project Aims

• Provide dosing guidance consistent with formulary
  standard of care,
• Examine patient pharmacotherapeutic indices
  relative to historical controls derived from the
  hospital data warehouse,
• Explore treatment diagnoses drug correlation
  in conjunction with utilization and
• Educate physicians on clinical pharmacologic
  principles specific to population and drug
  combinations of interest.

21
Pediatric Knowledgebase (PKB)Design Issues
Project Design Requirements Gathering
Project Scoping Charter, IRB Training
Steering Committee Formation, Prioritization
Design Team Physician champion for therapeutic
area, Clinical Pharmacologist / Modeler,
Programmer, IT specialist
Dashboard Prototype Development Forecasting DSS
Data Warehouse Access, Security, Modeling
PKB Shell SCM Interface User Interface Formulary
Metrics Questionnaire Clinical and operational
benefit
Steering Committee Clinical Care Attending
(Chair), Members IRB head, external
pharmacometrician, 3 physicians, project
sponsor, IT specialist, business manager,
hospital pharmacist
Testing
Presentation to Therapeutic Standards Committee
(TSC)
TSC Approval for production use granted by
Therapeutic Standards Committee
Refinement
Training and Implementation
22
Pediatric Knowledgebase (PKB)Design Issues -
Source Data
23
Pediatric Knowledgebase (PKB)Design Issues
Static Data
24
Pediatric Knowledgebase (PKB)Design Issues
Hospital Computing Environment
25
Methotrexate Dashboard

• Anti-folate chemotherapeutic agent
• Renal excretion
• Enterohepatic recirculation
• Toxicity at high or prolonged low exposure

26
Methotrexate Dashboard
27
Methotrexate Dashboard

• Dose?
• Dose adjustment?
• Therapeutic drug monitoring?
• Toxicity?

28
Methotrexate Dashboard

• 12 year-old boy with osteosarcoma and renal
  insufficiency.
• 3 year-old girl with leukemia and previous
  history of hyperbilirubinemia.

29
Methotrexate Dashboard

• Percentage of patients with elevated creatinine
  able to get full dose without toxicity.
• Most common toxicity in patients with elevated
  creatinine.

30
Methotrexate Dashboard
31
Methotrexate Dashboard

• Underlying model accounts for combined elements
  of methotrexate therapy
• Dose characteristics (amount, duration)
• Covariates (age, weight, gender, disease state,
  etc.)
• Pharmacokinetics (plasma concentration)
• Pharmacodynamics (creatinine clearance)
• Applied to individual patient data for TDM

32
Methotrexate Dashboard
Dose, infusion time
Central Compartment
Peripheral Compartment
Dissipation of Effect
Effect Compartment
Elimination from Plasma
33
Methotrexate Dashboard

• Current MTX data model
• Patients with normal renal function
• Patients with compromised renal function
• Very young patients (3 month to 1 year old)

34
Methotrexate Dashboard

• Provide predictions of
• MTX concentrations at later time
• Creatinine clearance at later time
• Time to reach threshold plasma concentration
• Guidance for dose titration
• Diagnosis of delayed MTX clearance due to acute
  nephrotoxicity
• Guidance of rescue therapy in response to renal
  toxicity

35
Methotrexate Dashboard
36
Methotrexate Dashboard
37
Methotrexate Dashboard
38
Methotrexate Dashboard
39
Methotrexate Dashboard
40
Methotrexate Dashboard
41
The PKB Team
Mahesh Narayan Sundarajaran Vijakumar,
PhD Kalpana Vijakumar Mark Schreiner, MD Rollie
Essex Arun Muralidharan Santhanam Srinivasa
Raghavan Theo Zaoutis, MD
Athena Zuppa, MD Jeffrey Skolnik, MD John
Mondick, PhD Kelly Wade, MD Peter C. Adamson,
MD Garret Brodeur, MD Manish Gupta, PhD Di Wu,
PhD Bhuvana Jayaraman Dimple Patel Dominique
Paccaly, PharmD
42
Questions?