Pfizer Global Research & DevelopmentSandwich CT13 9NJ Kent

Introduction

Sildenafil (Viagra) is an orally active, selective inhibitor of cGMP specific phosphodiesterase type 5 (PDE5), approved for the treatment of male erectile dysfunction.

Experimental evidence suggests that the nitric oxide-cGMP pathway may be important in producing clitoral engorgement, pelvic vasocongestion and vaginal lubrication thus enhancing the female sexual arousal response.

Although some clinical studies in FSAD have shown statistically significant sildenafil efficacy over placebo others have not.

Three Phase 2b/3 studies had collected data in a consistent manner thus offering a good opportunity for a combined analysis to look at temporal aspects of clinical response under placebo and active treatment.

Covariate Estimate SE CV

1 Intercept -0.428 0.237 55%2 Intercept -0.521 0.237 45%

3 Intercept -1.240 0.236 19%

4 Intercept -2.760 0.239 9%

5 Intercept -5.140 0.258 5%

6 Intercept -2.270 0.317 14%

7 Intercept -4.540 0.333 7%

8 Intercept -6.880 0.355 5%

9 Intercept -9.660 0.378 4%

10 Intercept 0.380 0.038 10%

11 Time 0.338 0.057 17%

12 drgtaken x dose x tlevel 0.001 0.001 66%

13 Age -0.035 0.005 13%

14 base.osat 0.821 0.059 7%

15 Drgtaken 1.010 0.088 9%

16 time exponent 0.096 0.028 29%

17 base.ssat 1.460 0.136 9%

20 Menopausal 0.003 0.001 36%

ln[j,org)] 0.272 0.039 14%

ln[SD j,sex)] 0.070 0.059 85%

Sexual Activity Daily Diary• Captured every day

• Sexual events– Presence or absence– Satisfaction scale (1-5) (Ssex)

– Event rated moderately to extremely satisfying (scale 3 to 5) were rated as satisfying

• Orgasm – Presence and absence – Satisfaction scale (1-5) (Sorg)

– Satisfying events defined as above

• Time of drug intake before sexual event– Less than 30’ minutes– 30’ to 4 hours– Greater then 4 hours

Modelling and Simulation of the Telephone Sexual Activity Daily Diary Data of patients

with female sexual arousal disorder treated with sildenafil (Viagra®)L. Claret(1), E.H. Cox(1), L. McFadyen(2), A. Pidgen(2), P.J. Johnson(2), S. Haughie(2), M. Boolell(2), R. Bruno(1)

(1)Pharsight Corporation, Drug Development Consulting Services, Mountain View, CA. (2)Pfizer, PGRD, Sandwich, UK

Randomized, double blind, placebo controlled, multicenter studies to evaluate efficacy, safety and toleration of oral sildenafil administered for 12 weeks to women with FSAD

Study 1127 - 248 women– Premenopausal (n = 43) and postmenopausal (n= 205) women on HRT– Minimum physiological level of estradiol ( 40 pg/ml) except for patients on HRT– Stratified according to free testosterone level ( 0.9 pg/ml (n = 121) or <0.9 pg/ml

(n = 127))– Dose: placebo (n = 124), 50 mg or adjusted doses (25 (6%) or 100 (75%) mg) (n =

124)

Study 1082 - 71 women– Postmenopausal on HRT with minimum physiological level of estradiol ( 40

pg/ml) and free testosterone ( 0.9 pg/ml) – Dose: Placebo, 5, 10, 25, 50, 100 mg with 21, 11, 9, 10, 10, 10 patients

respectively

Study 1123 - 98 women– Premenopausal with minimum physiological level of estradiol and free

testosterone– Dose: Placebo, 5, 10, 25, 50, 100 mg with 83, 41, 43, 42, 42, 43 patients

respectively

All studies have 2-6 weeks treatment free and 12 wks treatment phase.

Dose to be taken 30 mins to 4 hours prior to sexual event

Data

Study 1127Proportions of satisfying orgasms (Sorg≥3) versus time

(90% confidence interval)

Correlation of sexual satisfaction and orgasm satisfaction

Sexual satisfaction1 2 3 4 5

01

23

45

Org

asm

sat

isfa

ctio

n

Areas of tiles are proportional to the number of events

The structure of the model was determined by the nature of the clinical endpoint that is derived

from the TSADD

Sexual event

Orgasm0: no orgasm

1: not satisfied…

5: extremely satisfied

Covariates: Treatment duration (time)Orgasm satisfaction baseline score (base.Sorg)AgeDrug intake (drgtaken)Sildenafil Dose Testosterone level (tlevel)Menopausal status (meno)

Overall Sexual Satisfaction1: not satisfied

…5: extremely satisfied

Covariates: Treatment duration (time)sexual satisfaction baseline score (base.Ssex)

Time since previous event

Final Model

, 1 ,| expj org j org j orgP S m z g E θ , , 2 ,| , expj sex j j org sex j sexP S m z S g E θ

exp g x

1 expx

withx

1 1 5, , θ and 2 6 9, , θ , 0,j S N

16-13 14 15 12 20 11age+ base.Sorg+ drgtaken+ meno dose drgtaken tlevel 1-e time

orgE

16210 , 17 11base.Ssex 1 time

sex j orgE S e

• Time between events and scores are independent, time to event distribution was estimated by a Weibull distribution model in NONMEM:

• Orgasm, and sexual satisfaction scores were modeled simultaneously in NONMEM:

with for Weibull distribution t*

*

P( T t ) exp h( u )du , h t t

P( T t ) exp t ,

1

0

Sildenafil effect depends on drug intake, testosterone level, and

menopausal statusSexual satisfaction scores were modeled from predicted orgasm

satisfaction

Parameter estimates

Model Obj Parameter Value Ste Cv

expon. 59253 ln() -1.310 0.016 1%0.102 0.008 8%

Weibull 57485 ln() -1.410 0.016 1%ln() 0.316 0.012 4%

0.1250.008 7%

Time to event model Scores model

E t

1 11

where is the Gamma function

Posterior predictive check of P(Sorg N), and P(Ssex N) as a function of dose. We compared the 90%CI and median across 100 simulated replicates

of predicted proportions to those observed P(Sorg>=3)

rand

P(S

org

>=

3)

0 20 40 60 80 100

0.0

0.2

0.4

0.6

0.8

1.0

P(Sorg>=2)

rand

P(S

org

>=

2)

0 20 40 60 80 100

0.0

0.2

0.4

0.6

0.8

1.0

P(Ssex>=3)

rand

P(S

sex>

=3

)

0 20 40 60 80 100

0.0

0.2

0.4

0.6

0.8

1.0

P(Ssex>=2)

rand

P(S

sex>

=2

)

0 20 40 60 80 100

0.0

0.2

0.4

0.6

0.8

1.0

simul_scores3_post_rdtk3_meno.ssc

Dose (mg) Dose (mg)

Dose (mg)Dose (mg)

E(t) = 374 days

Simulation of orgasm satisfaction score • The probability of achieving a Orgasm Satisfaction Score, P(Sorg), 3 is used throughout the simulations.

• The response rate is evaluated for the expected patient population. –As such the performance of sildenafil in the expected patient population can be evaluated and is not affected by variability due to a limited trial size.

–Subsequent trial simulations can then be performed to evaluate how this population performance can be translated into the performance of a trial (of limited size).

• Simulations are integrated across model uncertainty:–500 replicates–90% prediction intervals are represented–A population of 5,000 subjects is simulated for each replicate

• Relevant patient covariates are sampled from the 1123/1127/1082 study population• Response rate is evaluated after 12 weeks of study duration (unless noted otherwise)

The expected absolute treatment effect for 100 mg Sildenafil may be 7%

1 median (90% prediction interval)

Sildenafil Dose (mg)

p(Sorg3) (%)

Absolute difference from placebo (%)

0 34.7 (31.3, 38.3) - 5 35.0 (31.6, 38.7) 0.3 (-0.2, 0.9)

10 35.4 (31.9, 39.0) 0.7 (0.1, 1.3)

25 36.4 (32.9, 40.1) 1.7 (0.8, 2.7)

50 38.1 (34.4, 42.0) 3.5 (1.9, 5.2)

100 41.6 (37.0, 46.3) 6.9 (3.7, 10.0)

Treatment responses for 100 mg Sildenafil may vary according to

testosterone level from 4.0% (1st quartile) to 12% (4th quartile).

0 20 40 60 80 100

02

46

81

0

Sildenafil Dose (mg)

P(S

org>

=3)

, ab

solu

te d

iffer

ence

fro

m p

lace

bo (

%))

0.1+ thru 0.9 pg/ml (Q1)0.9+ thru 1.2 pg/ml (Q2)1.2+ thru 1.6 pg/ml (Q3)1.6+ thru 18.4 pg/ml (Q4)

There is a substantial decline in orgasm satisfaction score with age. However, the impact of age on the

treatment effect is minimal.

P(S

org

>=

3)(

%)

20 30 40 50 60 70

02

04

06

08

0

100 mg sildenafilplacebo

P(S

org

>=

3),

ab

solu

te d

iffe

ren

ce f

rom

pla

ceb

o (

%))

20 30 40 50 60 70

05

10

15

Age (years)

100 mg Sildenafil

Conclusions• A model for the time-course of diary data observed in 3 Phase 2/3 studies of sildenafil was developed

• A Weibull distribution best described the probability density function of the time between sexual events• Satisfaction scores were simultaneously modelled with overall sexual satisfaction conditional on orgasm satisfaction

• Simulations were performed to evaluate the expected clinical response in the FSAD patient population• p(Sorg3) ranges from 34.7% for placebo to 41.6% for 100 mg sildenafil. Thus, the absolute treatment effect (difference from placebo) for sildenafil may be up to 6.9% for 100 mg sildenafil.• treatment effect of sildenafil is increased in post-menopausal women with high testosterone level.

Model

Model Simulations

Objectives

To develop models to:

• Characterize the probability of sexual events and their satisfaction scores over time based on Telephone Sexual Activity Daily Diary (TSADD) data obtained in clinical studies of sildenafil in patients with FSAD without concomitant Hypoactive Sexual Desire Disorder (HSDD).

• Simulate the expected dose-response in various patient populations to assess the impact of patient and disease characteristics on outcome.

p(O

rg. S

at. >

=3,

abs

olut

e di

ffere

nce

from

pla

cebo

(%

))

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake Testosterone level >= 0.9 pg/ml

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake Testosterone level < 0.9 pg/ml

PlaceboSildenafil

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake Testosterone level >= 0.9 pg/ml

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake Testosterone level < 0.9 pg/ml

PlaceboSildenafil

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake Testosterone level >= 0.9 pg/ml

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake Testosterone level < 0.9 pg/ml

PlaceboSildenafil

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake Testosterone level >= 0.9 pg/ml

Time week

Pro

ba

bili

ty

0 5 10

0.0

0.1

0.2

0.3

0.4

0.5

Satisfying orgasms with drug intake Testosterone level < 0.9 pg/ml

PlaceboSildenafil

… by menopausal status

Time week

Probability

0 5 10

0.0

0.2

0.4

0.6

0.8

1.0

Satisfying orgasms with drug intake Testosterone level >= 0.9 pg/ml

in post-menopausal

Time week

Probability

0 5 10

0.0

0.2

0.4

0.6

0.8

1.0

Satisfying orgasms with drug intake Testosterone level >= 0.9 pg/ml

in pre-menopausal

PlaceboSildenafil proptest_varia_1127_estim_DAAT.ssc

Time (week)