semi-mechanistic time-to-event modelling in malaria · pdf filesemi-mechanistic time-to-event...

Semi-Mechanistic Time-to-Event Modelling in Malaria - a pharmacometric approach to tropical medicine research

Joel Tarning

Associate Professor │ University of Oxford │ UK

Head of Clinical Pharmacology │ Mahidol-Oxford Tropical Medicine Research Unit │ Thailand

Joel Tarning, PhD

PAGE 2105 2 June 5, 2015 | Crete, Greece

Mahidol-Oxford Tropical Research Unit

“MORU aims to fight infectious tropical diseases affecting rural communities in Asia and elsewhere in the developing world. We develop effective and

practical means of diagnosing and treating Malaria and other neglected diseases such as Typhus, Melioidosis, and Leptospirosis“

(http://www.tropmedres.ac)

Joel Tarning, PhD


Neglected Tropical Diseases (NTDs)

NTDs thrive mainly among the poorest populations NTDs are endemic in 149 countries and affect more than 1.4 billion people NTDs cost developing economies billions of dollars every year

43% of the world’s pop. at risk of NTDs

1% of new entities from the public market

1% of the global R&D goes toward NTDs

Slingsby, JITMM, 2014

Joel Tarning, PhD


Malaria statistics (estimates) according to WHO

Half of the world’s population live in areas at risk of malaria transmission

An estimated 207 million clinical episodes and 627,000 deaths annually (2012)

Approximately 1,700 people die each day from malaria

Children under the age of five and pregnant women are most severely affected

Malaria is strongly associated with poverty

PK/PD properties are poorly understood for many of the current antimalarial drugs

Most antimalarial drugs were introduced at the wrong dose

Malaria

Joel Tarning, PhD


Malaria Malaria control and elimination mechanisms

Vector control - Indoor residual insecticide spraying - Long-lasting insecticide-treated bed nets - Availability, price, resistance, usage

Antimalarial drug therapy

- Artemisinin-based combination therapy (ACT); short acting drug + long acting drug

- Compliance, availability, price, resistance, drug quality

No available vaccines that can prevent malaria

Artemisinin-resistant malaria is the single greatest threat to our ability to control and eliminate malaria

Shrimp fishing with bed net (photo Moizo Bernard)

Dondorp et al, NEJM, 2009

Joel Tarning, PhD


PK/PD in children

Why a particular interest in paediatric patients? 85% of all malaria-related deaths occur in

children under the age of 5 Children are not small adults! Nonlinear relationship between body weight

(and age) and drug exposure

Anderson, Rev. Colomb. Anestesiol. 2013

Piperaquine: ↓ AUC [Tarning, 2012]

Artesunate/DHA: ↓/↓ AUC [Tarning, 2013]

Lumefantrine: ↓ AUC [Mwesigwa, 2010; Checchi, 2006]

Amodiaquine: ↓ AUC [Mwesigwa, 2010]

Chloroquine: ↓ AUC [Karunajeewa, 2008]

Sulfadoxine: ↓ AUC [Barnes, 2006]

Pyrimethamine: ↓ AUC [Barnes, 2006]

Artemether/DHA: ↑/↑ AUC [Mwesigwa, 2010]

Desethylamodiaquine: ↑ AUC [Mwesigwa, 2010]

2 years

Joel Tarning, PhD


Chotsiri et al, unpublished

Patients: – Intermittent seasonal preventive treatment in Burkina Faso – 183 children in PK/PD arm (0.2-5 years) – 562 children in PD arm (large efficacy trial) – High transmission season (August, September, October)

Drug regimen: – WHO guided dosing (2 mg/kg/day DHA and 18 mg/kg/day PQ) – 3 day fixed oral dose, once a month for three months

Sparse blood sampling: – Finger prick capillary sampling (~200 μl) – Pre-dose, 0-6 days, day 7, and 8-30 days

Piperaquine assay: – LC-MS/MS

NONMEM

Piperaquine PK/PD in children

Joel Tarning, PhD


T im e (d a y s )

PQ

co

nc

en

tra

tio

ns

(n

g/m

L)

0 2 0 4 0 6 0 8 0 1 0 0

1

1 0

1 0 0

1 0 0 0

A u g u s t

(6 3 C P /1 8 V P )

S e p te m b e r

(6 3 C P /1 8 V P )

O c to b e r

(6 3 C P /1 8 V P )

Venous Capillary

Observed piperaquine concentrations

Gut

Transit Vc

Vp1

Vp2

Sparse sampling 2-4 samples/individual PRIOR approach (Tarning et al, CPT, 2012) Allometric scaling of PK parameters 𝑃 𝛩 𝐷𝑎𝑡𝑎) =

𝑃 𝛩 ⋅ 𝑃 𝐷𝑎𝑡𝑎|𝛩

𝑃 𝐷𝑎𝑡𝑎

Pharmacokinetic model



Joel Tarning, PhD


L N P o p u la tio n P re d ic tio n s (n g /m L )

LN

Ob

se

rva

tio

ns

(n

g/m

L)

1 2 3 4 5 6 7

1

2

3

4

5

6

7

A )

L N In d iv id u a l P re d ic tio n s (n g /m L )

LN

Ob

se

rva

tio

ns

(n

g/m

L)

1 2 3 4 5 6 7

1

2

3

4

5

6

7

B )

L N P o p u la tio n p re d ic tio n s (n g /m L )

Co

nd

itio

na

l w

eig

hte

d r

es

idu

als

2 4 6

-4

-2

0

2

4C )

T im e a fte r d o s e (d a y s )

Co

nd

itio

na

l w

eig

hte

d r

es

idu

als

0 2 0 4 0 6 0

-4

-2

0

2

4D )

Visual predictive check

n = 2,000

Successfully stabilised PK with PRIOR approach

High predictive performance

=> PD modelling (time to infection)

3.76% (95%CI: 3.00-7.52%)

6.39% (95%CI: 2.82-7.33%)

Piperaquine PK/PD in children Basic Goodness-of-fit


Joel Tarning, PhD


0

50

100

150

200

250

300

350

400

450

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

1.E+10

0 1 2 3 4 5 6

Dru

g co

nce

ntr

atio

n (

ng/

mL)

Par

asit

e b

iom

ass

Time (weeks)

Malaria PK/PD PD modelling

Treatment of malaria

Sufficient drug concentrations to eliminate symptomatic infections?

Tota

l par

asit

e co

un

t

Limit of detection

Symptomatic infection

Post-prophylactic effect

Therapeutic success

MIC

Release of merozoites from the liver

MIC: Minimum Inhibitory Concentration

Joel Tarning, PhD


0

50

100

150

200

250

300

350

400

450

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

1.E+10

0 0.5 1 1.5 2 2.5 3 3.5 4

Dru

g co

nce

ntr

atio

n (

ng/

mL)

Par

asit

e b

iom

ass

Time (weeks)

Malaria PK/PD Preventive treatment

PD: Time to malaria infection

Sufficient drug concentrations to eliminate new infections?

Microscopic case detection / Lag-time

Tota

l par

asit

e co

un

t

Limit of detection Prevent new

infection

Post-prophylactic effect

MIC Lag-time

New infection

Time to malaria infection

Joel Tarning, PhD


T im e (d a y s )

Pa

ra

sit

e c

ou

nts

(/

L) T

ota

l bo

dy

pa

ra

site

1 0 -4

1 0 -2

1 0 0

1 0 2

1 0 4

1 0 6

1 0 2

1 0 4

1 0 6

1 0 8

1 0 1 0

1 0 1 2

E v e n t c e n s o r in g

in te rv a l fo r 1 06

p a ra s ite c o u n t ( / L )

E v e n t c e n s o r in g

in te rv a l fo r 1 02

p a ra s ite c o u n t ( / L )

-1 5 .4-2 4 .9 -5 .4-1 0 .6 0-20-30

Parasitemia corrected censored interval (Bergstrand et al, Science Trans Med, 2014)

Observed high parasitemia

Interval censoring TTE model

Hazard function:

𝐻𝑧 𝑡 = Θ𝐵𝐴𝑆𝐸 × 1 −𝐸𝑀𝐴𝑋 × 𝐶𝑃 𝑡

𝛾

𝐸𝐶50𝛾+ 𝐶𝑃 𝑡 𝛾

Survival function:

𝑆 𝑡 = exp − 𝐻𝑧 𝑡 𝑑𝑡𝑡

0

Slow growth rate: 5-fold every 48 hrs Fast growth rate: 10-fold every 48 hrs

Starting interval of erythrocyte infection


Probability of event occurring in the interval: Pr 𝑎 < 𝑡 < 𝑏 | Θ = 1 − 𝑆 𝑏 − 𝑆 𝑎 Mechanistic interpretation

of PK/PD estimates

Joel Tarning, PhD


Final PK/PD model PK: PRIOR – Allometry (BW) Parasite-based censoring

provided a more mechanistic interpretation

Constant hazard model with EMAX-type drug effect

Baseline hazard was 7.24 infections/year (CI: 4.61-11.5)

EC50 of piperaquine was 6.08 ng/mL (CI: 5.04-14.3)

Shape factor () was 1.46 (CI: 1.23-2.78)

Kaplan-Meier plot for event 3 (Run 3)

Time

Pe

rce

nta

ge

of su

rviv

al

50

60

70

80

90

100

500 1000 1500 2000 2500 3000

Visual predictive check

Observed time to new infections Prediction interval

Developed PK/PD model is suitable for simulations

n = 2,000



Joel Tarning, PhD


T im e (d a y s )

Su

rviv

al

pro

po

rtio

n

0 3 0 6 0

0 .5

0 .6

0 .7

0 .8

0 .9

1 .0

T im e (d a y s )

Su

rviv

al

pro

po

rtio

n

0 3 0 6 0

0 .5

0 .6

0 .7

0 .8

0 .9

1 .0

T im e (d a y s )

Su

rviv

al

pro

po

rtio

n

0 3 0 6 0 9 0 1 2 0

0 .5

0 .6

0 .7

0 .8

0 .9

1 .0

T im e (d a y s )

Su

rviv

al

pro

po

rtio

n

0 3 0 6 0 9 0 1 2 0

0 .5

0 .6

0 .7

0 .8

0 .9

1 .0

(A ) (B )

(C ) (D )

5-20 kg children 5-10 kg children

Increased dose (Tarning, CPT, 2012) Standard dose Reduced malaria

incident by 26% Reduced malaria incident by 34%

Reduced malaria incident by 27%

Reduced malaria incident by 19%



Single treatment Single treatment

Multiple treatments Multiple treatments

Increased dose (Tarning, CPT, 2012) Standard dose

Joel Tarning, PhD


Piperaquine PK/PD in children & adults

241 Karen and Burmese patients infected with Plasmodium vivax malaria

Standard 3-day treatment with oral dihydroartemisinin and piperaquine

Blood sampling from 116 patients at 6 random time points over 69 days and at relapsing malaria

LC-MS/MS

NONMEM

Tarning et al, CPT:PSP, 2014

Joel Tarning, PhD


Sparse data (simultaneous fit)

435 venous plasma samples

356 capillary plasma samples

Structural model

3-compartment disposition model

Flexible transit-absorption model

Concentration-time profile

Body-weight – allometric function



Joel Tarning, PhD



𝑅𝑒𝑙𝑎𝑝𝑠𝑒(𝑡) = 𝑆𝑢𝑟𝑔𝑒(𝑡)3𝑤𝑘 × 𝑆𝑢𝑟𝑔𝑒(𝑡)6𝑤𝑘 × 𝑆𝑢𝑟𝑔𝑒(𝑡)9𝑤𝑘

𝑆𝑢𝑟𝑔𝑒 𝑡 = 1 +𝐴𝑚𝑝𝑙𝑖𝑡𝑢𝑑𝑒

(𝑡 − 𝑇𝑖𝑚𝑒)2

𝑊𝑖𝑑𝑡ℎ2

𝛾

+ 1

Time (fix) Width

Amp γ

Vivax malaria in tropical regions displays frequent relapses in 3 week intervals (White, 2011)

Constant hazard function

A multiple surge function was implemented to characterize the periodically increased risk of relapsing malaria


Joel Tarning, PhD


Pharmacodynamics

parameters

Population estimates

[%RSE] 95% CI for estimates

θBHZ (relapses/year) 8.94 [22.1] 4.80-11.7

θSA (%) 123 [23.7] 61.8-165

θSW (h) 49.5 [54.3] 35.7-80.4

θSHP 4 Fixed -

PC50 (ng/ml) 6.92 [15.5] 6.15-11.3

γ 9.28 [25.6] 5.09-14.0

Drug effect and covariates were investigated

- Inhibitory maximum effect (EMAX) function - Covariate free PD model

PK-PD cohort: - 62 patients - Aged 11-56 years old - Weight 23-67 kg

Efficacy study: - 241 patients - Aged 1-57 years old - Weight 7-74 kg


ℎ 𝑡 = 𝜃𝐵𝐻𝑍 × 𝑟𝑒𝑙𝑎𝑝𝑠𝑒 𝑡 × 1 −𝐶(𝑡)𝛾

𝐶(𝑡)𝛾 − 𝐼𝐶50𝛾


Joel Tarning, PhD



Piperaquine fully suppressed the first relapse 3 weeks after the initial infection

Day 7 piperaquine concentration above 27 ng/mL can suppress the risk of relapse for 30 days

Major implications on the disease burden (morbidity of relapses)

99%


n=10,000

Joel Tarning, PhD


Half of the world’s population live in areas at risk of malaria transmission

Approximately 1,700 people die every day from malaria (mainly <5 year)

Malaria is a preventable and curable disease

We have used a pharmacometric approach to identify patient groups at particular risk of therapeutic failures (and resistance development)

We have used a pharmacometric approach to optimise the dosing in these sub-groups of patients, up to 34% reduction of malaria in young children

Dihydroartemisinin-piperaquine is a suitable treatment of vivax malaria, able to suppress fully the first relapse of malaria (morbidity impact)

We have shown that a pharmacometric approach is a highly useful tool in

tropical medicine research and can have a major impact on policy

Revised antimalarial treatment guidelines for young children

Concluding remarks

Joel Tarning, PhD


Acknowledgements

All patients that participated in these studies All Funders All colleagues in Burkina Faso All colleagues at MORU All colleagues at SMRU All colleagues at WWARN All colleagues at Oxford

Palang Chotsiri PhD student

Praiya Thana PhD student

Poster # I 54

Joel Tarning, PhD


Department of Clinical Pharmacology

Questions?

Thank you for your attention

Pharmacometricians Post-doc position available: [email protected]

semi-mechanistic time-to-event modelling in malaria · pdf filesemi-mechanistic time-to-event...

Documents