Title: Evaluation of Rapid and Sustained Population Viral ... - ACoP

illbreedinggrouseΠολεοδομικά Έργα

16 Νοε 2013 (πριν από 3 χρόνια και 9 μήνες)

69 εμφανίσεις

Title:
Evaluation of Rapid and Sustained Population Viral Response Rates Predicted Under Hepatitis C Viral Dynamic
Models

Authors:
Kyle Baron*(1), Patanjali Ravva (2), Vivek Purohit (2), Matthew M. Riggs (1), Marc R. Gastonguay (1)

Institutions:
(1) Metru
m Research Group, LLC, Tariffville, CT, USA; (2) Pfizer, New London, CT, USA.

Objectives:
A wide variety of mechanistic models describing hepatitis C virus (HCV) replication and the
pharmacodynamics of standard antiviral interventions currently exist in th
e published literature [1]. While regulators are
encouraging use of viral dynamic models to guide development of new antiviral therapies [2], the ability of viral
dynamic models to predict longitudinal population viral response after standard of care (SOC
) therapy [3] has not yet
been systematically evaluated. The objective of this study was to compare both short
-

and long
-
term population
-
level
viral responses predicted under two HCV dynamic models [4,5] with observed response rates in a meta data set comp
iled
from published clinical reports.

Methods:
HCV dynamic models published by Dahari et. al. [4] and Snoeck et. al. [5] were used to simulate individual
-
level viral load versus time profiles in
in silico

populations of HCV genotype
-
1 infected patients.

The Dahari model is
an extension of the model published by Neumann et. al. [6] and provides a theoretical basis for the observed diversity in
short
-
term viral load versus time profiles. The Snoeck model is a nonlinear mixed effects model estimated on a l
arge set
(2,100 patients, >21,000 observations) of long
-
term viral load versus time data taken from patients receiving SOC
treatment consisting of pegylated interferon
-
alfa
-
2a (IFN) and ribavirin (RBV). In the current simulation study, model
differential e
quations were written in fortran and solved using the LSODA solver called from the
deSolve()

R
package. Between
-
subject variability (BSV) in model parameters was generated by simulating log
-
transformed
individual
-
level parameters from univariate normal di
stributions in R with variance estimates as published by Snoeck et.
al. [5]. Viral load observations were simulated without residual unexplained variability. The simulated standard of care
intervention consisted of IFN 180 mcg/wk and RBV 13 mg/kg/d for 4
8 weeks and followed by a 24 week observation
period after withdrawing treatment. For both models, IFN and RBV potency was assumed to be as published in Snoeck
et. al. [5]. Viral response rates (viral load below quantitation limit [BQL], 50 copies/ml) we
re calculated at various times
after initiation of the intervention. Response milestones of interest included rapid viral response (RVR, undetectable
viral load at 4 weeks after starting treatment) and sustained viral response (SVR, undetectable viral loa
d 24 weeks after
treatment completion). A simple dropout model was implemented by labeling 0
-
40% of the
in silico
subjects as
“dropped” and assigning non
-
responder outcome without regard to the simulated profile [1]. For each scenario, 1000
subjects wer
e simulated. Simulated response rates from SOC treatment were compared to rates calculated on an
aggregate data set consisting of eleven clinical studies published from 2002 to 2010 involving similar long
-
term SOC
treatment in HCV genotype
-
1 infected pati
ents.

Results:
Approximately 5% (Snoeck model) to 10% (Dahari model) of simulated subjects receiving SOC treatment in
each run had a baseline viral load that was BQL and were excluded from the data set prior to response rate
determination. Individual vir
al load versus time profiles from the Snoeck and Dahari models were generally similar,
however the Dahari model was more likely to produce the “flat
-
partial” response [4]. Table 1 shows simulated response
rates for each model at various viral response mil
estones. Regardless of the dropout magnitude, the models under
-
predicted the RVR rate (6
-
8%) compared to the meta data set (weighted average: 15%, study
-
level range: 11
-
37%). The
simulated SVR rates without dropout were considerably higher (65
-
77%) than
the meta data estimate (weighted average:
48.7%, study
-
level range: 38
-
52%). Randomly dropping out 30% of the subjects resulted in simulated SVR rates (45
-
54%) that were comparable with that seen in the meta data set.

Conclusions:
SVR predictions from pub
lished HCV dynamic models [4,5] with no dropout mechanism were over
-
optimistic compared to rates seen in a similar SOC meta data set. Model
-
predicted SVR rates approached values
observed in the meta data set when typical dropout rates (~30%) were assumed.

Model
-
predicted RVR rates were about
one
-
half of the rate observed in the meta data set. Despite some qualitative differences in individual
-
level profiles, the
Snoeck and Dahari model parameters predicted similar population
-
level responses. The factors

identified here should be
explored in new model development before utilizing these models for guiding anti
-
HCV drug development




References:


[1] J Guedj, L Rong, H Dahari, and A S Perelson. A perspective on modelling hepatitis c virus infection. J Vira
l Hepat,
Aug 2010.


[2] Guidance for industry (DRAFT). Chronic Hepatitis C Virus Infection: Developing Direct
-
Acting Antiviral Agents
for Treatment. Center for Drug Evaluation and Research, United States Food and Drug Administration.
http://www.fda.gov/d
ownloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM225333.pdf.
Accessed January 17, 2010


[
3] Stephane Chevaliez and Jean
-
Michel Pawlotsky. Hepatitis c virus: virology, diagnosis and management of antiviral
therapy. World J Gastroenterol,
13(17):2461

6, May 2007.


[4] Harel Dahari, Arthur Lo, Ruy M Ribeiro, and Alan S Perelson. Modeling hepatitis c virus dynamics: liver
regeneration and critical drug efficacy. J Theor Biol, 247(2):371

81, Jul 2007.


[5] E Snoeck, P Chanu, M Lavielle, P Jacq
min, E N Jonsson, K Jorga, T Goggin, J Grippo, N L Jumbe, and N Frey. A
comprehensive hepatitis c viral kinetic model explaining cure. Clin Pharmacol Ther, 87(6):706

13, Jun 2010.


[6] A U Neumann, N P Lam, H Dahari, D R Gretch, T E Wiley, T J Layden, and
A S Perelson. Hepatitis c viral
dynamics in vivo and the antiviral efficacy of interferon
-
alpha therapy. Science, 282(5386):103

7, Oct

1998.



Table 1:

Simulation Results and Meta
-
Data Summary



Dropout Magnitude (%)

Meta Data Set

Week

Model

0

20

30

40

Weighted Rate

N Subjects

4

DH

6.39

5.25

4.85

3.19

15.1

1188

SN

8.23

6.34

5.07

5.38

12

DH

50.4

40.1

34.1

29.9

47.1

2223

SN

55.6

44.3

39.2

32.8

24

DH

76.3

61.5

53.7

45.6

66.9

598

SN

76.5

60.2

53.4

45.6

48

DH

83.9

67.4

58.4

49.6

64.6

1503

SN

82.1

65.7

58.5

48.9

72

DH

76.8

62.1

54.1

46.8

47.8

1801

SN

65

51.1

45.5

38.6

DH = Dahari et. al. model [4], SN = Snoeck et.al. model [5]