Download Document

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
Objectives
Introduction
Sildenafil (Viagra) is an orally active, selective inhibitor of cGMP specific phosphodiesterase type 5 (PDE5), approved for
the treatment of male erectile dysfunction.
To develop models to:
Although some clinical studies in FSAD have shown statistically significant sildenafil efficacy over placebo others have not.
• 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).
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.
• Simulate the expected dose-response in various patient
populations to assess the impact of patient and disease
characteristics on outcome.
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.
Data
Sexual Activity Daily Diary
• Captured every day
… by menopausal status
0.5
0.4
00
55
10
10
00
Time
Timeweek
week
55
Scores model
θ2  q6 ,
,q9  h j ,S  N  0, 
1.0
0.8
0.6
0.4
1.0
0.2
0.4
P(Sorg>=2)
0.6
0.8
1.0
0.0
0
20
40
Dose (mg)
60
80
100
0
20
rand
40
80
100
80
100
P(Ssex>=2)
1.0
P(Ssex>=3)
60
rand
Dose (mg)
0.8
E(t) = 374 days
where  is the Gamma function
0.6

P(Sorg>=2)
0.4
1
  1 

P(Sorg>=3)
0.8
1%
4%
7%
0.6
0.016
0.012
0.008
0.4
-1.410
0.316
0.125
P(Ssex>=2)
and
55%
45%
19%
9%
5%
14%
7%
5%
4%
10%
17%
66%
13%
7%
9%
29%
9%
36%
14%
85%
0.2
,q 5 
CV
0.237
0.237
0.236
0.239
0.258
0.317
0.333
0.355
0.378
0.038
0.057
0.001
0.005
0.059
0.088
0.028
0.136
0.001
0.039
0.059
0.0
θ1  q1 ,
SE
-0.428
-0.521
-1.240
-2.760
-5.140
-2.270
-4.540
-6.880
-9.660
0.380
0.338
0.001
-0.035
0.821
1.010
0.096
1.460
0.003
0.272
0.070
P(Sorg>=3)
ln()
ln()

Sexual satisfaction scores were
modeled from predicted orgasm
satisfaction
1  exp  x 
Estimate
Intercept
Intercept
Intercept
Intercept
Intercept
Intercept
Intercept
Intercept
Intercept
Intercept
Time
drgtaken x dose x tlevel
Age
base.osat
Drgtaken
time exponent
base.ssat
Menopausal
0.2
57485
Sildenafil effect depends on drug
intake, testosterone level, and
menopausal status
Covariate
0.0
Weibull
q1
q2
q3
q4
q5
q6
q7
q8
q9
q10
q11
q12
q13
q14
q15
q16
q17
q20
ln[hj,org)]
ln[SD hj,sex)]
1.0

0.2
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
0.8
1%
8%


0.0
10
10
Time
Timeweek
week
10
10
Cv
0.016
0.008
1
Probability
0.8
0.0
0.0
0.1
0.1
Areas of tiles are proportional to the number of events
P(Ssex>=3)
Ste
-1.310
0.102

0.6
54 3 2
1
0.5
0.5
0.3
0.3
0.2
0.2
0.3
0.3
Probability
Probability
0.4
0.4
Placebo
Placebo
Sildenafil
Sildenafil
0.2
Value
ln()

E t  
0.4
Probability
0.2
0.0
0.2
0.2
0.0
0.0
0.1
0.1
0.2
0.2
0.5
0.5
55
Time
Timeweek
week
59253

0
Orgasm satisfaction
0.5
0.5
0.4
0.4
0.3
0.3
0.3
0.3
Probability
Probability
0.1
0.1
0.0
0.0

exp  x 
0.2
0.2
Probability
Probability
0.1
0.1
10
10
P  S j ,sex  m | z j , S j ,org   g exp  θ2  Esex  h j , sex 
Eorg  q13  age+q14 base.Sorg+q15  drgtaken+ q12  q 20  meno   dose  drgtaken  tlevel  q11  1-e-q16 time

00
expon.
P( T *  t )  exp     t  ,
Esex  q10  S 2j ,org  q17  base.Ssex  q11  1  e q16 time
10
10
Parameter
5
0.3
Probability
Obj
4
Placebo
Placebo
Sildenafil
Sildenafil
Time to event model
Model
3
Satisfying
Satisfying orgasms
orgasms with
with drug
drug intake
intake
Testosterone
Testosterone level
level << 0.9
0.9 pg/ml
pg/ml
Time
Time week
week
t
 1
P( T *  t )  exp    h( u )du  , with for Weibull distribution h  t          t 
 0

2
Satisfying
Satisfyingorgasms
orgasms with
withdrug
drugintake
intake
Testosterone
Testosteronelevel
level << 0.9
0.9pg/ml
pg/ml
0.1
55
Time (week)
10
10
0.0
00
with g  x  
55
Parameter estimates
• Time between events and scores are independent, time to event distribution was
estimated by a Weibull distribution model in NONMEM:
10
0.2
0.5
0.4
0.3
0.2
0.1
Probability
1: not satisfied
…
5: extremely satisfied
Satisfying
Satisfyingorgasms
orgasms with
withdrug
drugintake
intake
Testosterone
Testosteronelevel
level >=
>= 0.9
0.9pg/ml
pg/ml
Time
Timeweek
week
0.0
Final Model
P  S j ,org  m | z j   g exp  θ1  Eorg  h j ,org 
Overall Sexual
Satisfaction
55
5
Tim e week
proptest_v aria_1127_estim_DAAT.ssc
Time
Timeweek
week
0.0
0.0
00
0
Placebo
Sildenaf il
0.0
Covariates:
Treatment duration (time)
sexual satisfaction baseline
score (base.Ssex)
10
10
Satisfying
Satisfying orgasms
orgasms with
with drug
drug intake
intake
Testosterone
Testosterone level
level >=
>= 0.9
0.9 pg/ml
pg/ml
• Orgasm, and sexual satisfaction scores were modeled simultaneously in NONMEM:
Orgasm
00
0.4
0.4
0.4
55
Time
Time week
week

0: no orgasm
1: not satisfied
…
5: extremely satisfied
0.3
0.1
0.0
00
Model
Covariates:
Treatment duration (time)
Orgasm satisfaction
baseline score (base.Sorg)
Age
Drug intake (drgtaken)
Sildenafil Dose
Testosterone level (tlevel)
Menopausal status (meno)
0.2
Probability
0.3
0.0
0.1
0.2
Probability
0.4
0.5
Satisfying
Satisfying orgasms
orgasms with
with drug
drug intake
intake
0.5
Satisfying
Satisfying orgasms
orgasms
Dose to be taken 30 mins to 4 hours prior to sexual event
Sexual event
10
10
Time
Timeweek
week
All studies have 2-6 weeks treatment free and 12 wks treatment phase.
Time since previous event
Probability
Probability
0.4
0.4
0.5
0.5
Satisfying
Satisfyingorgasms
orgasms with
withdrug
drugintake
intake
Study 1127
Proportions of satisfying orgasms (Sorg≥3) versus time
(90% confidence interval)
55
10
Sexual satisfaction
1
Satisfying
Satisfyingorgasms
orgasms
00
Satisfying orgasms with drug intake
Testosterone level >= 0.9 pg/ml
in pre-menopausal
Correlation of sexual satisfaction
and orgasm satisfaction
• Time of drug intake before sexual event
– Less than 30’ minutes
– 30’ to 4 hours
– Greater then 4 hours
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
5
Tim e week
• Orgasm
– Presence and absence
– Satisfaction scale (1-5) (Sorg)
– Satisfying events defined as above
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
The structure of the model was
determined by the nature of
the clinical endpoint that is
derived from the TSADD
0
0.6
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)
1.0
• Sexual events
– Presence or absence
– Satisfaction scale (1-5) (Ssex)
– Event rated moderately to extremely satisfying
(scale 3 to 5) were rated as satisfying
Satisfying orgasms with drug intake
Testosterone level >= 0.9 pg/ml
in post-menopausal
0.4
Randomized, double blind, placebo controlled, multicenter studies to
evaluate efficacy, safety and toleration of oral sildenafil administered for
12 weeks to women with FSAD
0
20
40
60
80
100
0
20
40
Dose (mg)
rand
60
rand
Dose
(mg)
simul_scores3_post_rdtk3_meno.ssc
Model Simulations
Absolute difference
from placebo (%)
0.3 (-0.2, 0.9)
0.7 (0.1, 1.3)
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)
0
5
10
1
15
80
5
60
40
6
10
10
8
P(Sorg>=3), absolute difference from placebo (%))
25
p(Sorg3)
(%)
34.7 (31.3, 38.3)
35.0 (31.6, 38.7)
35.4 (31.9, 39.0)
36.4 (32.9, 40.1)
Sildenafil Dose
(mg)
100 mg sildenafil
placebo
0
20
40
60
80
0
Sildenafil Dose (mg)
P(Sorg>=3)(%)
100
20
80
0
60
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)
4
40
100 mg Sildenafil
2
15
10
5
20
There is a substantial decline in
orgasm satisfaction score with age.
However, the impact of age on the
treatment effect is minimal.
0
–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)
0
Treatment responses for 100 mg
Sildenafil may vary according to
testosterone level from 4.0%
(1st quartile) to 12% (4th quartile).
P(Sorg>=3), absolute difference from placebo (%))
• Simulations are integrated across model uncertainty:
difference from placebo (%))
–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).
0
• 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.
The expected absolute treatment
effect for 100 mg Sildenafil may be 7%
p(Org. Sat. >=3,p(Org.
absolute
from placebo (%))
>=3, absolute
Sat. difference
Simulation of orgasm satisfaction score
100
20
Sildenafil Dose (mg)
30
40
50
60
70
20
30
Age (years)
median (90% prediction interval)
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(Sorg3) 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.
Pfizer Global Research & Development
Sandwich CT13 9NJ Kent
40
50
60
70