semi-mechanistic time-to-event modelling in malaria · pdf filesemi-mechanistic time-to-event...
TRANSCRIPT
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
PAGE 2105 3 June 5, 2015 | Crete, Greece
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
PAGE 2105 4 June 5, 2015 | Crete, Greece
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
PAGE 2105 5 June 5, 2015 | Crete, Greece
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
PAGE 2105 6 June 5, 2015 | Crete, Greece
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
PAGE 2105 7 June 5, 2015 | Crete, Greece
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
PAGE 2105 8 June 5, 2015 | Crete, Greece
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
Piperaquine PK/PD in children
Chotsiri et al, unpublished
Joel Tarning, PhD
PAGE 2105 9 June 5, 2015 | Crete, Greece
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
Chotsiri et al, unpublished
Joel Tarning, PhD
PAGE 2105 10 June 5, 2015 | Crete, Greece
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
PAGE 2105 11 June 5, 2015 | Crete, Greece
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
PAGE 2105 12 June 5, 2015 | Crete, Greece
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
Piperaquine PK/PD in children
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
PAGE 2105 13 June 5, 2015 | Crete, Greece
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
Piperaquine PK/PD in children
Chotsiri et al, unpublished
Joel Tarning, PhD
PAGE 2105 14 June 5, 2015 | Crete, Greece
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%
Piperaquine PK/PD in children
Chotsiri et al, unpublished
Single treatment Single treatment
Multiple treatments Multiple treatments
Increased dose (Tarning, CPT, 2012) Standard dose
Joel Tarning, PhD
PAGE 2105 15 June 5, 2015 | Crete, Greece
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
PAGE 2105 16 June 5, 2015 | Crete, Greece
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
Piperaquine PK/PD in children & adults
Tarning et al, CPT:PSP, 2014
Joel Tarning, PhD
PAGE 2105 17 June 5, 2015 | Crete, Greece
Piperaquine PK/PD in children & adults
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
Tarning et al, CPT:PSP, 2014
Joel Tarning, PhD
PAGE 2105 18 June 5, 2015 | Crete, Greece
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
Piperaquine PK/PD in children & adults
ℎ 𝑡 = 𝜃𝐵𝐻𝑍 × 𝑟𝑒𝑙𝑎𝑝𝑠𝑒 𝑡 × 1 −𝐶(𝑡)𝛾
𝐶(𝑡)𝛾 − 𝐼𝐶50𝛾
Tarning et al, CPT:PSP, 2014
Joel Tarning, PhD
PAGE 2105 19 June 5, 2015 | Crete, Greece
Piperaquine PK/PD in children & adults
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%
Tarning et al, CPT:PSP, 2014
n=10,000
Joel Tarning, PhD
PAGE 2105 20 June 5, 2015 | Crete, Greece
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
PAGE 2105 21 June 5, 2015 | Crete, Greece
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
PAGE 2105 22 June 5, 2015 | Crete, Greece
Department of Clinical Pharmacology
Questions?
Thank you for your attention
Pharmacometricians Post-doc position available: [email protected]