PK/PD analysis in assessment of abuse deterrence

Megan J. Shram, PhD

Director and Principal, Altreos Research Partners Inc.

Adjunct Professor and Lecturer, Department of Pharmacology, University of TorontoINC Research Consultant

ADF Science Meeting October 30 – September 1, 2013

Bethesda, MD

Acknowledgments

• Consultant to pharmaceutical and biotech companies

• Thank you to Purdue Pharma, LP for permission to present data on reformulated OxyContin® and OXN

Outline

• Background on PK/PD in abuse potential

• What is PK/PD relationship of opioids?

– Determinants of variation

• Application of PK/PD analysis in development of abuse-deterrent formulations (ADF)

Introduction

• Draft Guidance on Assessment of Abuse Potential of Drugs (January 2010):– “Characterization of the PK/PD properties of a…product is

important for determining the abuse potential of a…product.”

• Draft ADF guidance (January 2013):– “PK data should be collected to correlate with the PD

outcomes.”

– “The rate of rise of drug onset for the intact and manipulated potentially abuse-deterrent formulation should be given appropriate weight in the overall analysis of the abuse deterrent properties.”

Introduction

• Goal of PK/PD analysis is to assist in predicting the effect of a drug over time, in relation to exposure

• In terms of abuse, the relationship between rate of rise of [drug] and effect is considered important– Fast onset, short duration of action favors repeated self-administration

– Delaying onset, extending duration of action can reduce immediate reinforcing effect and need to take more drug

+++

--

+

++

Questions

• Can PK assist in predicting subjective response to opioid?

• How can one assess if change in PK profile will be enough to affect PD response?

• eg, partial defeat of the AD mechanism

System

• Age

• Gender

• Opioid experience

• Tolerance/Dependence

• Expectations

• Genetic variations

Pharmacodynamics(CV ≈50- >100%)

• Subjective effects• Miosis• Analgesia• Respiratory depression

• Adverse events

• Behavior

Pharmacokinetics(CV ≈30%)

• Opioid, Dose

• Formulation, Route

• BBB permeability

• ADME, protein binding

• Active metabolites

• Plasma vs. Effect site

• Arterial v venous sampling/timing

PK/PD: Determinants of Variation

??EndpointsAnalysis

Interpretation

CV=coefficient of variation

Assessing the link between opioid exposure

and results of a clinical abuse potential study

Recent ADF Examples

When PK/PD relationship might exist

• Physicochemical barrier– Resistance to manipulation for oral, IN and IV*

administration

• Prodrug– Must be cleaved systemically to liberate active moiety

– Deterrence from IN and IV administration, and possibly oral if limited by saturable process

• Delivery System– Modified drug delivery to reduce diversion via

multiple routes

*not generally feasible to administer manipulated product IV

PK/PD of intranasally administered Reformulated OxyContin®

• 5-way crossover study in recreational opioid users with intranasal experience

– Coarse and fine ground OTR (30 mg)

– Positive controls: Oxy API and fine ground original OC (30 mg)

– Placebo control: lactose powder

Time course of effects

0

5

10

15

20

25

30

35

40

45

50

0 4 8 12 16 20 24

Pla

sma

oxy

cod

on

e (n

g/m

L)

Time post-dose (hr)

Oxycodone concentration

OTR fine

OTR coarse

OC

0

10

20

30

40

50

60

70

80

90

100

0 4 8 12 16 20 24D

rug

Liki

ng

VA

S (0

-10

0)

Time post-dose (hr)

Drug Liking VAS

OTR fine OTR coarse OC

Effects unrelated to drug exposure can impact experience (at the moment and overall)

Courtesy of Purdue Pharma; Perrino et al., 2012 CPDD; Harris et al., submitted

Rate of Rise: “Abuse quotient”

• Lower Cmax and longer Tmax lower Abuse Quotient

0

20

40

60

80

100

120

140

160

180

200

Cmax/Tmax

OTR fine

OTR coarse

OC

Oxy API

0

20

40

60

80

100

120

140

160

180

200

Emax/TEmax

OTR fine

OTR coarse

OC

Oxy API

Drug Liking Visual Analog ScaleOxycodone Pharmacokinetics

Data courtesy of Purdue Pharma

Exposure-Response Relationship

• For reformulated OxyContin®, delaying and lowering Cmax had significant impact on liking

• But if PK “drives” PD, what is concentration-effect relationship?

PK/PD Correlations: Subjective Effects

y = 0.5386x + 54.143R² = 0.0518

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100

Dru

g Li

kin

g V

AS

(0-1

00)

Plasma oxycodone concentration (ng/mL)

OTR finey = 0.6612x + 44.435

R² = 0.1077

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100

Dru

g Li

king

VA

S (0

-100

)

Plasma oxycodone concentration (ng/mL)

OTR coarse

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100

Emax

of

Dru

g Li

kin

g V

AS

(0-1

00)

Cmax of oxycodone (ng/mL)

Overall R2=0.12y = 0.2782x + 72.519

R² = 0.0602

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100

Dru

g Li

king

VA

S (0

-100

)

Plasma oxycodone concentration (ng/mL)

OC fine

Derived Parameters By Timepoint

Data courtesy of Purdue Pharma

PK/PD Correlations: Physiological Effects

y = -0.0509x + 5.3132R² = 0.1486

0

1

2

3

4

5

6

7

8

0 20 40 60 80 100

Pup

il si

ze (

mm

)

Plasma oxycodone concentration (ng/mL)

OTR fine

y = -0.0586x + 5.7401R² = 0.3608

0

1

2

3

4

5

6

7

8

0 20 40 60 80 100

Pupi

l siz

e (m

m)

Plasma oxycodone concentration (ng/mL)

OTR coarse

y = -0.0249x + 4.4345R² = 0.2132

0

1

2

3

4

5

6

7

8

0 20 40 60 80 100

Pupi

l siz

e (m

m)

Plasma oxycodone concentration (ng/mL)

OC fine

0

1

2

3

4

5

6

7

8

0 20 40 60 80 100

MP

C (

mm

)

Cmax of oxycodone (ng/mL)

Overall R2=0.27

Derived Parameters By Timepoint

Data courtesy of Purdue Pharma

When PK/PD might matter less

• Opioid agonist-antagonist combinations– Antagonist attenuates/reverses effect of opioid

– Does PK of antagonist help predict response to agonist?

• Aversion– Aversive agent not intended to impact exposure to

agonist

– PK would NOT predict overall response to ADF

– Might still experience high from opioid; however, product not liked

PK/PD of intranasally administered OXN

• 3-way crossover study in recreational opioid users with intranasal experience

– Crushed OXN 40/20 mg (oxycodone/naloxone)

– Positive control: Oxy API 40 mg

– Placebo control: lactose powder

Time course of effects

Data courtesy of Purdue Pharma

0

10

20

30

40

50

60

70

80

90

100

0 4 8 12 16 20 24

Pla

sma

oxy

cod

on

e (n

g/m

L)

Time post-dose (hours)

Plasma Concentration

Naloxone (ng/mL)

Oxy API - Oxycodone

OXN - Oxycodone

• At 2:1 ratio, OXN significantly reduces Drug Liking to placebo-like levels

• Rate of rise: “Abuse Quotient” (Cmax/Tmax) not applicable to opioid agonist-antagonist combinations, such as OXN

0

10

20

30

40

50

60

70

80

90

100

0 4 8 12 16 20 24

Dru

g Li

kin

g V

AS

(0-1

00

)

Time post-dose (hours)

Drug Liking VAS

OXN

Oxy API

Placebo

Antagonist PK – Agonist PD

• Highly variable concentrations and lack of concentration-effect relationship of to-be-marketed formulation

• Relationship between antagonist PK and agonist PD typically evaluated earlier in development in dose-ranging trials – ie, dose-effect

Data courtesy of Purdue Pharma

y = 0.15x + 47.987R² = 0.008

0

10

20

30

40

50

60

70

80

90

100

0 5 10 15 20 25 30 35

Dru

g Li

kin

g V

AS

(0-1

00

)

Plasma naloxone concentration (ng/mL)

Conclusions

• Relationship between PK and PD of abuse potential is weak and highly variable

• Effects unrelated to opioid exposure impact subject’s experience

• Clinical PD study necessary to determine potential for abuse (or its deterrence)

• PK alone cannot be used as substitute in abuse potential assessment