EMAEMA--EFPIA M&S Workshop EFPIA M&S Workshop --

Session 3Session 3 M&S examples that failed or succeeded to M&S examples that failed or succeeded to

meet regulatorsmeet regulators’’

expectationsexpectations

Decisive support of Modeling & Simulation Decisive support of Modeling & Simulation for getting drug approval of nonfor getting drug approval of non--tested tested dosing scheme : 2 examplesdosing scheme : 2 examples

Valérie CossonF. Hoffmann–La Roche Ltd

Position statement and Position statement and associated questions associated questions

Successful approval of non-tested dosing scheme using M&S techniques without further dedicated prospective studies

Would in general the EMA accept the principle of relying on M&S approaches to label an unstudied dose or dosing regimen?◦

If yes, what information and evidence would be needed by the EMA?

Would in general the EMA accept to increase the dose in a sub-population if the exposure remain within the investigated range? ◦

If yes, what information and evidence would be needed by the EMA?

2

BackgroundBackground

What has triggered the changes in dose (and dose adjustment rules) compared to the original plan?

Changes of regulatory recommendation

on endpoint target range: the example of C.E.R.A

(Continuous

Erythropoietin Receptor Activator), a new erythropoietin stimulating agent (ESA) for the treatment of anemia in patients with chronic kidney disease (CKD).◦

BLA and MAA were submitted to both FDA and EMA based on an hemoglobin (Hb) target range of 11 to 13 g/dL.

◦

After submission, FDA modified the Hb target range to 11 to 12 g/dL (as optimal target range) and avoid Hb > 13 g/dL for safety concern.

3

BackgroundBackground

What has triggered the changes in dose (and dose adjustment rules) compared to the original plan?

Less efficacy in an identified sub-population: the example of ribavirin

in HCV genotype-1 patients

with normal transaminases (ALT)◦

A lower dose than the recommended one for HCV genotype-1 infected patients with elevated ALT was used in patients with normal ALT

◦

Less efficacy i.e. sustained virologic response

(SVR) was

observed in this population compared to the patients with elevated ALT

4

Rationale and objectives of the M&S Rationale and objectives of the M&S analyses analyses

In both examples, the knowledge of the exposure- response relationship was considered sufficient to rely on

M&S approaches to investigate new doses

Clinical trial simulations were conducted to assess efficacy and safety clinical outcomes of non-tested dosing scheme (and dose adjustment) proposed in the SmPC (Summary of Products Characteristics)

5

Available Data for C.E.R.A.Available Data for C.E.R.A.

Data from 400 patients of 3 open-label, randomized, multicentre, parallel-group Phase III studies, AMICUS, MAXIMA and PROTOS

The PK assessment period in AMICUS was the 24-week correction period. The PK assessment period in both MAXIMA and PROTOS was the 28-week dose titration period.

The Hb assessments were performed weekly in the three Phase III studies during the correction or titration periods.

6

Available Data for RibavirinAvailable Data for Ribavirin

For model development, SVR and Hb data from genotype- 1 infected CHC patients with elevated ALT levels

receiving a ribavirin treatment of 48 weeks were used: ◦

817 patients for GAM development of SVR ◦

1233 patients for GAM development of incidence of anaemia

For the assessment of dose in genotype-1 infected patients with normal ALT levels receiving a daily dose of 800 mg ribavirin for 48 weeks, SVR data from 138 patients and Hb data from 206 patients were used

7

MethodsMethods

Development of PKPD models using available data.

Evaluation of the predictive performance of the PKPD models by visual predictive check and posterior predictive check on defined metrics

Simulations to evaluate the efficacy and safety of alternative dosing scheme.

8

The ExposureThe Exposure--Response Analysis for Response Analysis for C.E.R.A.C.E.R.A.

Time [days]

C.E

.R.A

. [ng

/mL]

05

1015

20

0 50 100 150 200

Hem

oglo

bin

[g/d

L]6

810

1214

16

PROTOS: patient 2651

Time [days]

C.E

.R.A

. [ng

/mL]

05

1015

20

0 50 100 150 200

Hem

oglo

bin

[g/d

L]6

810

1214

16

PROTOS: patient 752

Time [days]

C.E

.R.A

. [ng

/mL]

05

1015

20

0 50 100 150 200

Hem

oglo

bin

[g/d

L]6

810

1214

16

PROTOS: patient 851

Time [days]

C.E

.R.A

. [ng

/mL]

010

3050

0 50 100 150 200

Hem

oglo

bin

[g/d

L]6

810

1214

16MAXIMA: patient 1364

Time [days]

C.E

.R.A

. [ng

/mL]

010

3050

0 50 100 150 200

Hem

oglo

bin

[g/d

L]6

810

1214

16

MAXIMA: patient 4004

Time [days]

C.E

.R.A

. [ng

/mL]

010

3050

0 50 100 150 200

Hem

oglo

bin

[g/d

L]6

810

1214

16

MAXIMA: patient 2501

Time [days]

C.E

.R.A

. [ng

/mL]

010

2030

40

0 50 100 150 200

Hem

oglo

bin

[g/d

L]6

810

1214

16

AMICUS: patient 301

Time [days]

C.E

.R.A

. [ng

/mL]

010

2030

40

0 50 100 150 200

Hem

oglo

bin

[g/d

L]6

810

1214

16

AMICUS: patient 834

Time [days]

C.E

.R.A

. [ng

/mL]

010

2030

40

0 50 100 150 200

Hem

oglo

bin

[g/d

L]6

810

1214

16

AMICUS: patient 125

Observed and predicted Hb concentrations in individual patients selected from PROTOS, MAXIMA and AMICUS.

9

The ExposureThe Exposure--Response Analysis for Response Analysis for RibavirinRibavirin

The higher incidence of anemia in patients with normal ALT is due to a difference in percentage of female patients (61% for normal ALT versus 33% for elevated ALT)

Dose/BW (mg/kg)

Pro

babi

lity

SV

R

5 10 15 20 25

0.0

0.2

0.4

0.6

0.8

1.0

Probability SVR

Dose/BW (mg/kg)

Like

lihoo

d H

gB <

10

5 10 15 20 25

0.0

0.2

0.4

0.6

0.8

1.0

Likelihood AnemiaLikelihood of Anemia

1.0

0.8

0.6

0.4

0.2

0.0

Inci

denc

e H

b <

10 g

/dL

25 252015105

Probability SVR

Dose/BW (mg/kg)

0.0

0.4

0.2

0.8

0.6

1.0

Pro

babi

lity

SVR

5 2010 15Dose/BW (mg/kg)

Elevated ALTNormal ALT

10

Model qualificationModel qualification

The PKPD models are qualified to be used in simulation mode

30 35 40 45 50 55 60

050

100

150

200

250

SVR response (%)

Rep

licat

ions

800 mg

30 35 40 45 50 55 60

050

100

150

200

250

SVR response (%)

Rep

licat

ions

1000/1200 mg

0 5 10 15 20

050

100

150

200

250

Incidence HgB < 10 (%)

Rep

licat

ions

800 mg

0 5 10 15 20

050

100

150

200

250

Incidence HgB < 10 (%)

Rep

licat

ions

1000/1200 mg

Efficacy

Safety

Patients with elevated ALTPatients with normal ALT

Observation

Occurrence Hb>13 g/dL (IV)

C.E.R.A. Ribavirin

11

40 60 80

Patients with at least one Hb>13 during first 28 w [%]

AMICUS: 71

05

1015

2025

30

Num

ber o

f rep

licat

es

M&S Results: Simulation of efficacy and M&S Results: Simulation of efficacy and safety of nonsafety of non--tested dosing scheme and dose tested dosing scheme and dose adjustment rules of C.E.R.A.adjustment rules of C.E.R.A.

With 0.3 µg/kg/w

IV or SC every 2 weeks, the median predicted occurrence of Hb>13 g/dL

was decreased down to 41% and 26.5% respectively compared to 71% in AMICUS.

With 0.3 µg/kg/w

IV and SC given every 2 weeks, the median response rate (percentage of patients with Hb11 g/dL

and Hb

from baseline 1 g/dL

at least once during the first 24 weeks of treatment) was 92.5% and 83% respectively. The value in AMICUS was 93%.

0.3 µg/kg/w

IV 0.3 µg/kg/w

SC

Median predicted incidence >13 g/dL 41.0 % 26.5%

Median simulated response rate 92.5 % 83.0%

12

M&S Results: Simulation of efficacy and safety of M&S Results: Simulation of efficacy and safety of nonnon--tested dosing scheme of tested dosing scheme of RibavirinRibavirin

At 1000/1200 mg, the SVR in patients with normal ALT is predicted to be similar to patients with elevated ALT

800 mg 1000/1200 mg

Elevated ALT 40% (36%-45%) 49% (46%-53%)

Normal ALT 39% (34%-44%) 48% (42%-53%)

At 1000/1200 mg/day, the incidence of anaemia in patients with normal ALT is predicted to be higher than in patients with elevated ALT, especially in females

1000/1200 mg

Females Males

Elevated ALT

23% (19%-27%)

8% (6%-10%)

Normal ALT

31% (24%-38%)

11% (7%-16%)

13

ConclusionsConclusions

1.

The trial simulations have shown that new dosing scheme proposed in SmPC

ensures good efficacy and

manageable safety risk.

2.

Thanks to Modeling and Simulation techniques additional confirmatory trials could be avoided.

14

Regulatory FeedbackRegulatory Feedback

The new dosing scheme and dose adjustment for C.E.R.A. were approved by EMA and FDA

EMA accepted the changes in the label◦

The dose of 1000/1200 mg Ribavirin

is not limited only

to patients with elevated transaminases◦

The risk of developing anemia is higher in the female population

15

Position statement and Position statement and associated questions associated questions

Successful approval of non-tested dosing scheme using M&S techniques without further dedicated prospective studies

Would in general the EMA accept the principle of relying on M&S approaches to label an unstudied dose or dosing regimen?◦

If yes, what information and evidence would be needed by the EMA?

Would in general the EMA accept to increase the dose in a sub-population if the exposure remain within the investigated range? ◦

If yes, what information and evidence would be needed by the EMA?

16

Extended QuestionsExtended Questions

If the Sponsor demonstrates that:◦

The pharmacology and the mechanism of action of the drug are well known,

◦

The exposure-response relationships for efficacy and safety in the target population are adequately characterized,

◦

The covariate effects are known,◦

Enough confidence in the models to simulate response in a specific sub-population within the range of previously studied exposure:

All of the proof of the model robustness using classical technique (gof

plots, SE, VPC, PPC…) are provided

In which cases ◦

The EMA would only rely on the simulations to label an unstudied

dose or dosing regimen ?◦

The EMA would still ask to confirm in an additional study ?

17

BOS4 : Extended QuestionsBOS4 : Extended Questions

Significant Benefit in life-threatening disease

Prevention or delay of long term disease with high morbidity mortality

Symptomatic treatment

AEs

that can be monitored and treated

M&S only? M&S only? M&S only?

Safety signals with clear predictive indicators

M&S only? M&S only? M&S only?

Safety signals without clear predictive indicators

M&S

+

Additional data?

M&S

+

Additional data?

??

In which cases ◦

The EMA would only rely on the simulations to label an unstudied dose or dosing regimen ?

◦

The EMA would still ask to confirm in an additional study ?

18

Efficacy

Safety