Duloxetine versus placebo in the treatment of

stress urinary incontinence

Peggy A. Norton, MD,a Norman R. Zinner, MD,b Ilker Yalcin, PhD,c and Richard C. Bump, MD,c
for the Duloxetine Urinary Incontinence Study Group
Salt Lake City, Utah, Torrance, Calif, and Indianapolis, Ind

OBJECTIVE: The purpose of this study was to assess the efficacy and safety of duloxetine, a selective in-
hibitor of serotonin and norepinephrine reuptake, in the treatment of stress urinary incontinence.
STUDY DESIGN: A double-blind, randomized, placebo-controlled study was conducted in 553 women aged
18 to 65 years with a predominant symptom of stress urinary incontinence. Subjects were randomized to
placebo (n = 138 women) or duloxetine at one of three doses (20 mg/d, n = 138 women; 40 mg/d, n = 137
women; or 80 mg/d, n = 140 women). Outcome variables that were assessed after 12 weeks of treatment in-
cluded incontinence episode frequency recorded in a real-time diary and answers provided to the Patient
Global Impression of Improvement scale and the Incontinence Quality of Life questionnaire.
RESULTS: Duloxetine was associated with significant and dose-dependent decreases in incontinence
episode frequency that paralleled improvements that were observed in the Patient Global Impression of Im-
provement scale and the Incontinence Quality of Life questionnaire. The median incontinence episode fre-
quency decrease with the use of the pooled diary analysis with placebo was 41% compared with 54% for
duloxetine 20 mg per day (P = .06), 59% for duloxetine 40 mg per day (P = .002), and 64% for duloxetine 80
mg per day (P < .001). One half of the subjects at the 80 mg per day dose had a ≥64% reduction in inconti-
nence episode frequency (P < .001 vs placebo); 67% had ≥50% reduction (P = .001 vs placebo). These im-
provements were observed despite significant concurrent dose-dependent increases in the average voiding
interval in the duloxetine groups compared with the placebo group. Similar statistically significant improve-
ments were demonstrated in a subgroup of 163 subjects who had more severe stress urinary incontinence
(≥14 incontinence episode frequency per week; 49%-64% reduction in incontinence episode frequency in the
duloxetine groups compared with 30% in the placebo group). Discontinuation rates for adverse events were
5% for placebo and 9%, 12%, and 15% for duloxetine 20, 40, and 80 mg per day, respectively (P = .04).
Nausea was the most common symptom that led to discontinuation. None of the adverse events that were
reported were considered to be clinically severe.
CONCLUSION: This trial provides evidence for the efficacy and safety of duloxetine as a pharmacologic
agent for the treatment of stress urinary incontinence. (Am J Obstet Gynecol 2002;187:40-8.)

Key words: Stress urinary incontinence, pharmacologic treatment, duloxetine, quality of life, nor-
epinephrine, serotonin reuptake inhibitor

Numerous studies have implicated the neurotransmit-
ters serotonin and norepinephrine in the central neural
control of lower urinary tract function. Serotonergic ago-
nists generally suppress parasympathetic activity and en-
From the Department of Obstetrics and Gynecology, University of Utah,a
Doctor’s Urology Group,b and Lilly Research Laboratories, Lilly Corpo-
rate Center.c
Supported by Eli Lilly and Company.
Presented at the Twenty-Second Annual Meeting of the American Urogy-
necologic Society, Chicago, Ill, October 25-28, 2001.
The members of the Duloxetine Urinary Incontinence Study Group are
listed in the Appendix.
Reprint requests: Richard C. Bump, MD, Eli Lilly and Company Corpo-
rate Center, DC 2200, Indianapolis, IN 46285. E-mail: bump_richard@
lilly.com
© 2002, Mosby, Inc. All rights reserved.
0002-9378/2002 $35.00 + 0 6/6/124840
doi:10.1067/mob.2002.124840
hance sympathetic and somatic activity in the lower uri-
nary tract, which are effects that promote urine storage;
antagonists have the opposite effects.1-3 Electrical stimu-
lation of the medullary raphe nuclei that contain the cell
bodies of bulbospinal serotonergic neurons has been
shown to inhibit reflex bladder contraction in cats,4 and
there seem to be multiple sites of serotonergic modula-
tion in the spinal cord.1 Noradrenergic agonists and an-
tagonists also produce effects on sympathetic and somatic
activity in the lower urinary tract, which is dependent on
the adrenergic receptor subtype with which the agonists
and antagonists interact.5-10
On the basis of these potential continence-promoting
properties of serotonin and norepinephrine, animal
studies have been conducted with duloxetine hy-
drochloride, a combined norepinephrine and serotonin

40
Volume 187, Number 1
Am J Obstet Gynecol
reuptake inhibitor, to determine its effects on lower uri-
nary tract function. One study indicated that duloxetine
significantly increased bladder capacity and sphincteric
muscle activity in the cat acetic acid model of irritated
bladder function.11 The results also showed that the ef-
fects of duloxetine on the bladder and sphincter were
mediated centrally through both motor efferent and
sensory afferent modulation. These central effects of
duloxetine on the bladder were reversed by the nonse-
lective serotonergic antagonist methiothepin; the cen-
tral effects on the sphincter were blocked by prazosin
and LY53857, α1-adrenergic, and 5HT2 serotonergic re-
ceptor antagonists, respectively. Thus, it is logical to as-
sume that the effects on bladder detrusor and urethral
skeletal muscle activity were mediated through tempo-
ral prolongation of serotonin and norepinephrine in
the synaptic cleft.
The ability of duloxetine to increase activity of the stri-
ated urethral sphincter suggests that the compound
could have a role in the treatment of stress urinary in-
continence (SUI). This possibility was examined in a
double-blind, randomized placebo-controlled study of
140 subjects who had SUI with 20-mg, 30-mg, and 40-mg
once-daily doses of duloxetine during a 6-week treatment
period.12 When all the duloxetine arms were pooled, sta-
tistically significant improvements were seen in the fre-
quency of incontinent episodes and the 1-hour stress pad
test (SPT), but not in the 24-hour pad weight test. How-
ever, a dose-response relationship was not observed. Al-
though the results of this double-blind study were
encouraging, several study-design limitations were identi-
fied. First, a relatively narrow range of duloxetine doses
was evaluated. Second, marked intrasubject variability
was observed in the 1-hour SPT weight, the primary effi-
cacy measurement, which led to poor reproducibility. Fi-
nally, the study did not have specific criteria for the
diagnosis of stress incontinence.
The aim of this phase II study was to compare the ef-
fects of duloxetine (20, 40, or 80 mg/day) and placebo in
the treatment of SUI, as measured by incontinence
episode frequency (IEF) and condition-specific quality-
of-life measures. A wide range of duloxetine doses, a
twice-daily dosing schedule, and a treatment period of 12
weeks were selected to evaluate for a dose-response rela-
tionship. Finally, a more stringent standardized diagnos-
tic algorithm was established to obtain a study population
more likely to have genuine stress incontinence as the
predominant cause of the incontinence.
Material and methods
Incontinent female outpatients aged 18 to 65 years
who had a clinical diagnosis of SUI for at least 3 months
in duration were invited to participate in this double-
blind, placebo controlled, randomized clinical trial. To
enroll, subjects must have reported a predominant symp-
Norton et al 41
Table I. PGI-I scale
Check the one number that best describes how you have felt
overall since you began receiving this medication. (Psychiatric
stem)
1. Very much better
2. Much better
3. A little better
4. No change
5. A little worse
6. Much worse
7. Very much worse
Check the one number that best describes how your urinary
tract condition is now, compared to how it was before you
began receiving medication in this study. (Lower urinary tract
stem)
1. Very much better
2. Much better
3. A little better
4. No change
5. A little worse
6. Much worse
7. Very much worse
tom of SUI with ≥4 incontinent episodes per week, where
an episode was defined as an easily noticeable leakage of
urine that wets a pad or clothing and occurs with a phys-
ical stress such as coughing, sneezing, or exercising.
Additional requirements included urinary diurnal fre-
quency ≤7 per day, nocturnal frequency ≤2 per day, the
absence of predominant symptoms of enuresis or urge
incontinence, and no previous continence or prolapse
surgical procedure. All women underwent a supine, 400-
mL saline solution bladder infusion with a funnel at a
rate of 100 mL per minute without pressure measure-
ments. Subjects who were unable to tolerate the filling,
who had a first sensation of bladder filling at <100 mL, or
who had no sensation at any time during the filling were
excluded. After bladder filling, both a positive cough
stress test (CST, visualization of urine leakage concur-
rent with a cough) and SPT (leakage of >2.0 g) were re-
quired for inclusion. A subgroup of 86 women had
multichannel urodynamics, of whom 79 women (92%)
had confirmed genuine stress incontinence. The study
was conducted at 48 study centers in the United States.
The institutional review board for each site approved the
study and written informed consent was obtained from
all participants.
After a 2-week no-drug lead-in period followed by a
2-week placebo lead-in period, subjects were randomized
under double-blind conditions to 12 weeks of treatment
with duloxetine 20 mg/day (20 mg, once daily), duloxe-
tine 40 mg/day (20 mg, twice daily), duloxetine 80
mg/day (40 mg, twice daily), or placebo. All subjects re-
ceived two identical capsules twice daily. Subjects were
seen at 4-week intervals throughout the treatment pe-
riod. A final visit occurred 2 to 4 weeks after discontinua-
tion of duloxetine, although all subjects were taking
blinded placebo (Fig 1, A).
42 Norton et al July 2002
Am J Obstet Gynecol

Table II. Baseline clinical characteristics of illness severity (by recall and diary)

Characteristic Duloxetine 20 mg/d Duloxetine 40 mg/d Duloxetine 80 mg/d Placebo

Randomized patients (No.) 138 137 140 138

White (%) 93 96 91 94
Age (y) 49.4 ± 7.3 49.4 ± 8.0 49.3 ± 8.8 50.2 ± 8.9
Height (cm) 163.8 ± 6.8 164.0 ± 7.1 163.8 ± 5.9 162.8 ± 6.9
Weight (kg) 81.2 ± 21.8 77.8 ± 16.5 76.2 ± 14.5 78.7 ± 18.0
Body mass index (kg/m2) 30.3 ± 8.1 28.9 ± 6.2 28.5 ± 5.6 29.8 ± 7.2
By baseline recall
Micturition episodes per 24 h 6.3 ± 1.4 6.8 ± 1.5 6.7 ± 1.5 6.4 ± 1.5
Incontinent episodes per 24 h 2.5 ± 1.9 2.2 ± 1.8 2.5 ± 1.8 2.4 ± 1.7
By baseline diary
Micturition episodes per 24 h 9.2 ± 2.6 10.0 ± 2.6 9.8 ± 2.5 9.4 ± 2.5
Incontinent episodes per 24 h 1.6 ± 1.3 1.7 ± 1.6 1.9 ± 1.6 1.6 ± 1.1

Data are given as mean ± SD, unless otherwise indicated.

A computerized voice response system at a central lo-
cation was used to control randomization. To prevent
potential imbalance at baseline in both treatment assign-
ment and incontinence severity, a sequential treatment
assignment (dynamic allocation) was performed with the
use of an algorithm similar to the one outlined by Pocok
and Simon.13 For balancing in terms of incontinence
severity, stratification was performed based on IEF at
baseline (group I, >0 and <10 episodes per week; group
II, ≥10 and <20 episodes per week; group III, ≥20
episodes per week). Each subject was assigned to one of
the four treatment groups so that the numbers of subjects
who were assigned to treatments were balanced at each
investigative site and so that the treatment groups were
balanced within the IEF stratum across all investigators.
The primary efficacy measure was the IEF, recorded
real-time on daily diaries for 1 week before each baseline
and treatment visit. The daily diaries were also used to
collect the number of voids, the time of voids, and the
time that the study medication was ingested. Two diaries
were completed before randomization: one diary during
the second week of the no-drug lead-in and 1 diary dur-
ing the second week of the placebo lead-in. Three diaries
were completed during the active treatment phase of the
study; each diary was completed during the week before
the 3 monthly visits. One diary was obtained during the
week before the final study visit, although all subjects
were receiving blinded placebo at that time (Fig 1, A).
One secondary measure of efficacy was the Inconti-
nence Quality of Life (I-QOL) questionnaire.14,15 This is
a validated disease-specific 22-item instrument that evalu-
ates the effects of urinary incontinence in three domains
(avoidance and limiting behavior, social embarrassment,
and psychosocial impact). It includes questions that eval-
uate both the distress and impact of urinary inconti-
nence, with a score of 100 being the best possible and a
score of 0 being the worst possible quality of life. A sec-
ond quality-of-life measure, the Patient Global Impres-
sion of Improvement scale (PGI-I), has been validated as
an assessment tool in psychopharmacologic research, but
with a different stem qualifier.16 It is reproduced in Table
I with both its psychiatric and urinary tract stems; re-
sponse options were identical for both conditions.
Other efficacy measures included the fixed bladder
volume (400 mL) CST, followed by a fixed volume SPT
that included the performance of 10 vigorous coughs, 10
deep knee bends, and walking up and down the equiva-
lent of 50 steps. Cure was defined according to three cri-
teria that included a negative fixed volume CST (400
mL), a negative SPT (≤ 2 g), or no incontinent episodes
on the last urinary diary.
Safety was assessed by the evaluation of treatment-
emergent adverse events, discontinuations for adverse
events, vital sign measurements, and clinical laboratory
tests. Adverse events were elicited by nonprobing inquiry
at each visit and were recorded regardless of perceived
causality. An event was considered treatment emergent if
it occurred for the first time or worsened during the dou-
ble-blind treatment period.
Subject compliance with the protocol was assessed at
each visit by a count of the unused medication and by the
monitoring of the diaries. A subject was considered non-
compliant if she missed more than four consecutive doses
of medication and/or did not ingest at least 80% of her
medication. A subject was also considered noncompliant
if she did not return at least three satisfactory daily diaries
for each of the weeks before the treatment visits. Subjects
were not discontinued for noncompliance.
Analysis of variance (ANOVA) models were used to eval-
uate ranked percent changes in IEF and SPT, ranked
changes in the number of voids and average voiding in-
terval, and raw changes in quality of life and its domains.
The ANOVA model included treatment, site (as a fixed ef-
fect), and treatment-by-site interaction. If the treatment-
by-site interaction did not show a trend toward statistical
significance (P < .10), then the treatment-by-site term was
Volume 187, Number 1
Am J Obstet Gynecol
Fig 1. Study design and comparison of the last diary analysis (A)
and the pooled diary analysis (B) for the percent decrease in IEF.
The groups of vertical lines denote the weeks that the diaries were
completed.
dropped from the model. Investigative sites with fewer
than two randomized subjects per treatment group were
pooled for statistical analysis purposes. For pairwise com-
parisons of each duloxetine arm to placebo, least squares
means were used from the appropriate ANOVA model.
Analysis of percent change in IEF was performed with the
last visit diary for both the baseline and treatment phases
(Fig 1, A) and with pooled data from both baseline diaries
and all three treatment phase diaries (Fig 1, B).
Categoric variables (including PGI-I, CST, and treat-
ment-emergent adverse events) were analyzed by the
Pearson χ2 test. Enrollment into the study was set to end
when approximately 500 patients (125 per treatment
group) had been assigned to a treatment group at visit 3.
The sample size was determined to provide at least 80%
power to detect a pair-wise treatment difference to
placebo of 20% in the mean percent change in weekly
IEF with an overall type I error of 0.05 with the use of the
Dunnett multiple comparison procedure.
All analyses were based on the intent-to-treat principle
and included all randomized subjects with at least one post-
randomization outcome measure. Analyses were per-
formed with SAS software (SAS Institute Inc, Cary, NC). A
two-sided α level of .05 was considered statistically signifi-
cant for treatment effects.
Results
Originally 1124 female subjects were screened for eligi-
bility; 553 women met entry criteria and were randomized
to either duloxetine 20 mg/day (n = 138 women), dulox-
etine 40 mg/day (n = 137 women), duloxetine 80 mg/day
(n = 140 women), or placebo (n = 138 women). Ninety-six
percent of the women (132/138) receiving placebo, 96%
of the women (132/138) receiving duloxetine 20 mg/day,
94% of the women (129/137) receiving duloxetine 40
mg/day, and 93% of the women (130/140) receiving du-
Norton et al 43
Fig 2. Percent change in IEF by treatment assignment and visit.
Visit 4 was after 4 weeks of treatment, visit 5 was after 8 weeks of
treatment, and visit 6 was after 12 weeks of treatment. D20, Du-
loxetine 20 mg/day; D40, duloxetine 40 mg/day; D80, duloxe-
tine 80 mg/day.
loxetine 80 mg/day had at least one postrandomization
outcome measure and were included in the intention-to-
treat analysis. Demographic data and baseline clinical and
baseline diary data are noted in Table II. None of the data
differed significantly among the four treatment groups.
Treatment compliance, within the limits set by the pro-
tocol, was 78% in the duloxetine groups and 83% in the
placebo group. The mean exposure to the study drug was
74 days, and the median exposure was 84 days of a possible
90 days for all subjects who received duloxetine treatment.
The placebo response rate was a 40% reduction in in-
continence episodes by both the last diary and pooled
diary analysis (Table III). The last visit analysis showed sig-
nificant, but not dose-dependent, reductions at 40- and
80-mg/day doses of duloxetine compared with placebo
(Table III). In contrast, analysis of the pooled diary data
revealed significant and dose-dependent responses to du-
loxetine. These dose-dependent responses in IEF differ-
ences paralleled the improvements observed in the PGI-I
and I-QOL (Table III). One half of the subjects at the 80-
mg dose level had at least a 64% reduction in IEF
(P < .001 vs placebo), and 67% of the subjects had at least
a 50% reduction (P = .001 vs placebo). The improve-
ments in IEF in the duloxetine groups were observed, de-
spite statistically significant (compared with placebo)
decreases in the number of voids per week and increases
in the mean voiding interval (Table III). Improvements
in IEF with duloxetine were significant at the first treat-
ment phase visit after 4 weeks (Fig 2). At the end of the
final placebo washout phase of the study, IEF rebounded
in all duloxetine groups (increasing by 12% to 15%), but
not in the placebo group (decreasing by 9%).
Subgroup analysis of the group of subjects with more se-
vere SUI (arbitrarily defined before analysis as IEF of ≥14
44 Norton et al July 2002
Am J Obstet Gynecol

Table III. Efficacy assessments for all evaluable subjects

Duloxetine 20 mg/d Duloxetine 40 mg/d Duloxetine 80 mg/d Placebo

Randomized patients (No.) 138 137 140 138

Median IEF (%)* –44 [.6]† –59 [.02]† –58 [.04]† –40
Median IEF (%)‡ –54 [.06]† –59 [.002]† –64 [<.001]† –41
PGI-I (%)§ 31 [.5] 37 [.09] 44 [.005] 27
I-QOL mean change +5.3 [.6]† +7.8 [.16]† +9.3 [.03]† 5.8
Voids/day (No.)* –0.8 [.4]† –1.0 [.09]† –0.8 [.3]† –0.5
Voids/day (No.)‡ –1.0 [.05]† –1.2 [.003]† –1.4 [<.001]† –0.6
Mean void interval (min)* +16 [.05]† +18 [.02]† +18 [.006]† 5
Mean void interval (min)‡ +16 [.004]† +19 [<.001]† +24 [<.001]† 7

*Changes with the last visit diary analysis.

†Significance [P value] versus placebo (ANOVA with treatment and site as effects).
‡Changes with the pooled diary analysis.
§Percentage in “very much better” or “much better” categories.
Significance [P value] versus placebo (Pearson’s χ2 test).

Table IV. Efficacy assessments for subjects with IEF ≥14 at baseline

Duloxetine 20 mg/d Duloxetine 40 mg/d Duloxetine 80 mg/d Placebo

Randomized patients (No.) 39 37 45 42

IEF median (%)* –35 [.5]† –61 [.02]† –58 [.045]† –30
IEF median (%)‡ –49 [.2]† –63 [.004]† –64 [<.001]† –30
PGI-I (%)§ 21 [.6] 41 [.17] 50 [.02] 26
I-QOL mean total change 3.1 [.4]† 9.4 [.1]† 13.3 [.005]† 4.1
Avoidance/limiting behavior¶ 2.9 [.5]† 10.4 [.15]† 14.6 [.01]† 5.7
Psychosocial¶ 3.1 [.5]† 7.0 [.3]† 11.3 [.01]† 3.1
Social embarrassment¶ 3.4 [.5]† 11.9 [.03]† 15.0 [.01]† 3.2

*Changes with the use of the last visit diary analysis.

†Significance [P value] versus placebo (ANOVA with treatment and site as effects).
‡Changes with the use of the pooled diary analysis.
§Percent in “very much better” or “much better” categories.
¶Changes in the 3 I-QOL domains.
Significance [P value] versus placebo (Pearson’s χ2 test).

per week on pooled baseline diaries; Table IV) showed vir-
tually identical dose-dependent decreases in IEF with du-
loxetine; the placebo response decreased from 41% to
30%. There were corresponding improvements in PGI-I
and I-QOL scores and in all I-QOL subscales at the 80-
mg/day dose. Again a rebound in IEF was observed in the
duloxetine groups (+6% to +9%), but not for the placebo
group (–19%) during the washout period.
There was large variability in SPT changes in all treat-
ment groups, and no significant differences were noted.
The median SPT percent changes from baseline were
–30% (–100% to +2175%), –11% (–100% to +3240%),
–43% (–100% to +5800%), and –29% (–100% to
+12,333% [ie, from a loss of 0.30 g at visit 3 to 37.30 g at
visit 6]) with placebo and duloxetine 20 mg per day, 40
mg per day, and 80 mg per day groups, respectively. Diary,
SPT, and CST cure rates also did not differ significantly
among groups (Table V).
Discontinuation rates for adverse events were 5% for
the placebo group and 9%, 12%, and 15%, respectively,
for the duloxetine 20 mg, 40 mg, and 80 mg groups
(P = .04 overall; Table VI). The percentages of subjects
who reported ≥1 treatment-emergent adverse event are
noted in Table VII. Nausea was the most common adverse
event. Twelve of 18 women (two thirds) who had nausea at
the 80 mg per day dose continued the medication, despite
this side effect. The overall discontinuation rate for nau-
sea at the 80 mg per day dose was 4% (6/140). None of
the adverse events that were reported were considered to
be clinically severe. Five subjects had adverse events that
required hospitalization (one event before randomiza-
tion, one event while ingesting duloxetine 20 mg per day,
two events while ingesting duloxetine 40 mg per day, and
one event while ingesting duloxetine 80 mg per day); only
one of these events (rash) was felt to be related to the
study drug.
Comment
The results of this study provide evidence for the effi-
cacy of duloxetine as a pharmacologic agent for the treat-
ment of SUI, with significant improvements in both IEF
and quality of life. The subgroup analysis of the more se-
verely incontinent subjects (IEF ≥14 at baseline) indi-
cated that improvements that were observed with
Volume 187, Number 1 Norton et al 45
Am J Obstet Gynecol

Table V. Cure rates

Duloxetine 20 mg/d Duloxetine 40 mg/d Duloxetine 80 mg/d Placebo

% No. % No. % No. % No.

SPT ≤2 g* 32.7 110 43.2 111 38.9 113 36.8 114

Negative CST 9.8 112 22.3 112 15.8 114 12.7 118
Zero incontinent episodes of diary 16.4 128 24.4 123 18.7 123 15.2 132

No significant difference for any cure rates, duloxetine versus placebo.

*At 1 hour.

Table VI. Discontinuation rates because of adverse events*

Duloxetine 20 mg/d (No.) Duloxetine 40 mg/d (No.) Duloxetine 80 mg/d (No.) Placebo (No.)

No. 138 137 140 138

Discontinuations† 13 (9%) 17 (12%) 21 (15%) 7 (5%)
Nausea 2 5 6 1
Headache 1 1 2 0
Somnolence 1 0 2 0
Dizziness 0 2 1 0
Menorrhagia 2 0 0 0

*Only adverse events with >2 in a treatment arm are listed.

†P = .04, for overall treatment effect.

Table VII. Treatment-emergent adverse events reported during treatment*

Duloxetine 20 mg/d Duloxetine 40 mg/d Duloxetine 80 mg/d Placebo

No. 138 137 140 138

At least one adverse event (%) 62 68 73 61
Headache (%) 7 10 8 9
Nausea (%)† 9 9 13 2
Diarrhea (%) 5 4 4 3
Constipation (%) 4 4 6 1
Dry mouth (%) 4 5 7 1
Dizziness (%) 2 6 7 2
Insomnia (%)† 2 7 7 1
Sinusitis (%) 4 4 4 6
Fatigue (%)† 1 8 10 3
Nasopharyngitis (%) 8 4 6 4
Upper respiratory tract infection (%) 2 2 1 5

*All occurred in ≥5% of subjects in any treatment arm.

†P < .05, for overall treatment effect (Pearson’s χ2 test).

duloxetine treatment were maintained, despite increased
baseline incontinence severity.
The significant dose-dependent response to duloxe-
tine in the current trial differs from earlier findings12
that did not encounter a dose-dependent response for
either efficacy or safety. However, that earlier study eval-
uated a relatively narrow range of duloxetine doses (20
mg/d, 30 mg/d, and 40 mg/d) during a 6-week treat-
ment period; the present study assessed a broader
range of doses during a 12-week treatment period. In
addition, the earlier protocol did not specify any crite-
ria for the diagnosis of incontinence type. In the pre-
sent study, more stringent diagnostic criteria were
defined, which resulted in a more homogeneous group
of women, most of whom were likely to have genuine
stress incontinence.
The placebo response rate in this study was apprecia-
ble, likely related to the intensity of the protocol and the
frequency of diary completion. Nonetheless, significant
decreases in IEF with duloxetine 40 and 80 mg per day
compared with placebo were noted both with analysis
with the use of data from the last baseline and treatment
visit diaries and with analysis with the use of data from
pooled baseline and treatment diaries; the net improve-
ment (that is, placebo percent reduction subtracted
from the duloxetine percent reduction) varied between
18% and 23% for duloxetine 40 and 80 mg per day, de-
pending on the analytic method. For the subgroup of
46 Norton et al
more severely incontinent women, this net improvement
for duloxetine 40 and 80 mg per day was 10% higher
(28%-34%) because of the decreased placebo response
in this subgroup. Only the pooled analysis demonstrated
a dose response that paralleled the dose response ob-
served with the PGI-I and I-QOL instruments. This differ-
ence in dose responsivity is a consequence of the
considerable intrasubject variability in IEF that was ob-
served among individual diaries. Because SUI occurs with
physical activity, a large intrasubject variability in IEF, es-
pecially for subjects with mild and moderate SUI, may re-
sult from weekly variations in physically stressful activity.
By pooling diaries, intrasubject variability in IEF was re-
duced by approximately 50%, which allowed detection of
the same dose-response relationship in the decrease in
IEF that was observed with the quality-of-life end points.
Because quality-of-life scales reflect the entire treatment
phase experience of a subject, it is not surprising that her
pooled diaries (which also reflect the entire treatment
phase) are more reflective of her experience than a sin-
gle-point diary. Another way to overcome the large intra-
subject variability effect of single diary analysis is to
perform much larger clinical trials. This is being done in
ongoing multinational trials.
Many women find that it is possible to decrease their
frequency of incontinent episodes by voiding more fre-
quently, thus keeping their bladders less full. This
learned behavior may be reinforced and encouraged as a
result of the completion of multiple urinary diaries in a
clinical trial setting. Thus, it was important to be certain
that improvements that were observed with duloxetine
were not the result of increased micturition frequency. In
fact, just the opposite was observed, with significantly
fewer voids and with significantly increased time between
voids being observed in all duloxetine groups compared
with placebo.
The significant improvements in quality of life that
were demonstrated in this trial are important because
they help put the issue of improvement versus cure into
clinical perspective. Although relatively few subjects in
this trial had an elimination of their incontinence, at the
80 mg/day dose many women had a measurable im-
provement in their quality of life. Thus, for many sub-
jects, the median 64% reduction in IEF had a positive
impact on their lives.
It is noted often that urinary incontinence is not a
threat to life expectancy, but rather a threat to quality life
expectancy. As a consequence, quality of life has been in-
creasingly promoted as an outcome domain that should
be assessed in incontinence treatment trials, especially
for treatments (such as medications or behavioral inter-
ventions) that are more likely to improve than totally
cure incontinence. Several committees that have been
charged with the establishment of standards for out-
comes research that is related to urinary incontinence
July 2002
Am J Obstet Gynecol
and other pelvic floor disorders have specifically recom-
mended that the domain of quality of life be assessed with
the use of condition-specific instruments.17-21 The I-
QOL14,15 that was used in the current study has demon-
strated psychometric properties of validity, reliability, and
responsiveness and is one of several quality-of-life ques-
tionnaires that were recommended by the Symptom and
Quality of Life Assessment Committee of the 1998 World
Health Organization that sponsored the International
Consultation on Incontinence.17
Duloxetine was generally well tolerated, although dis-
continuation rates for adverse events were statistically
higher among subjects who received duloxetine therapy
than among subjects who received placebo therapy. The
duloxetine discontinuation rates of 9% to 15% were 4%
to 10% higher than those rates reported with placebo.
These differences in discontinuation rates are compara-
ble to those rates that have been cited for phenyl-
propanolamine (4.3%) or phenylpropanolamine plus
conjugated estrogen (8%),22 which are agents that were
used sometimes for the pharmacologic treatment of
stress incontinence before the Food and Drug Adminis-
tration’s ban on phenylpropanolamine because of its as-
sociation with hemorrhagic stroke.23 They are also
comparable to discontinuation rates that have been re-
ported for medications that are used to treat symptoms of
bladder overactivity (3%-45%).22,24-26
In summary, this study provides evidence for the effi-
cacy and safety of duloxetine in the treatment of SUI.
Large multinational randomized, blinded placebo-con-
trolled clinical trials to further document efficacy and tol-
erability are currently ongoing.
We thank Stephanie C. Koke, MS, for her assistance in
the preparation of this manuscript.
REFERENCES
1. Espey MJ, Du H-J, Downie JW. Serotonergic modulation of
spinal ascending activity and sacral reflex activity evoked by
pelvic nerve stimulation in cats. Brain Res 1998;798:101-8.
2. Thor KB, Hisamitsu T, de Groat WC. Unmasking of a neonatal
somatovesical reflex in adult cats by the serotonin autoreceptor
agonist 5-methoxy-N,N-dimethyltryptamine. Brain Res Dev
Brain Res 1990;54:35-42.
3. Danuser H, Thor KB. Spinal 5-HT2 receptor-mediated facilita-
tion of pudendal nerve reflexes in the anaesthetized cat. Br J
Pharmacol 1996;118:150-4.
4. MacMahon SB, Spillane K. Brain stem influences on the
parasympathetic supply to the urinary bladder of the cat. Brain
Res 1982;234:237-49.
5. Downie JW, Bialik GJ. Evidence for a spinal site of action of
clonidine on somatic and viscerosomatic reflex activity evoked
on the pudendal nerve in cats. J Pharmacol Exp Ther
1988;246:352-8.
6. Downie JW, Espey MJ, Gajewski JB. Alpha 2-adrenoceptors not
imidazole receptors mediate depression of a sacral spinal reflex
in the cat. Eur J Pharmacol 1991;195:301-4.
7. Espey MJ, Downie JW, Fine A. Effect of 5-HT receptor and
adrenoceptor antagonists on micturition in conscious cats. Eur J
Pharmacol 1992;221:167-70.
Volume 187, Number 1
Am J Obstet Gynecol
8. Gajewski J, Downie JW, Awad SA. Experimental evidence for a
central nervous system site of action in the effect of alpha-adren-
ergic blockers on the external urinary sphincter. J Urol
1984;132:403-9.
9. Krier J, Thor KB, de Groat WC. Effects of clonidine on the lum-
bar sympathetic pathways to the large intestine and urinary blad-
der of the cat. Eur J Pharmacol 1979;59:47-53.
10. Pederson E, Torring J, Klemar B. Effect of the alpha-adrenergic
blocking agent thymoxamine on the neurogenic bladder and
urethra. Acta Neurol Scand 1980;61:107-14.
11. Thor KB, Katofiasc MA. Effects of duloxetine, a combined sero-
tonin and norepinephrine reuptake inhibitor, on central neural
control of lower urinary tract function in the chloralose-anes-
thetized female cat. J Pharmacol Exp Ther 1995;274:1014-24.
12. Zinner N, Thor KB, Yalcin I, DeBrota D, Faries D, Riedl P. Effi-
cacy and safety of duloxetine in stress urinary incontinent pa-
tients. Proceedings of the 19th Annual Meeting of Urodynamics
Society; 1998 May 30; San Diego, Calif. San Diego: The Society;
1998.
13. Pocok S, Simon R. Sequential treatment assignment with bal-
ancing of prognostic factors in controlled clinical trials. Biomet-
rics 1975;31:103-15.
14. Wagner TH, Patrick DL, Bavendam TG, Martin ML, Buesching DP.
Quality of life of persons with urinary incontinence: develop-
ment of a new measure. Urology 1996;47:67-71.
15. Patrick DL, Martin ML, Bushnell DM, Yalcin I, Wagner TH,
Buesching DP. Quality of life of women with urinary inconti-
nence: further development of the incontinence quality of life
instrument (I-QOL). Urology 1999;53:71-6.
16. Guy W. ECDEU assessment manual for psychopharmacology.
Rockville (MD): National Institute of Mental Health, US Depart-
ment of Health, Education, and Welfare; 1976. p. 218-22.
17. Donovan J, Naughton M, Gotoh M, Corcos J, Jackson S, Kelleher C,
et al. Symptom and quality of life assessment. In: Abrams P,
Shoury S, Wein A, editors. Incontinence. Plymouth (UK):
Health Publication Ltd; 1999. p. 295-332.
18. Mattiasson A, Peters T, Schafer W, Blaivas J, Jonas U, Stanton S,
et al. Research methodology in incontinence. In: Abrams P,
Shoury S, Wein A, editors. Incontinence. Plymouth (UK):
Health Publication Ltd; 1999. p. 893-929.
19. Mattiasson A, Djurhuus JC, Fonda D, Lose G, Nordling J, Stohrer
M. Standardization of outcome studies in patients with lower uri-
nary tract dysfunction: a report on general principles from the
Standardization Committee of the International Continence So-
ciety. Neururol Urodyn 1998;17:249-53.
20. Lose G, Fantl JA, Victor A, Walter S, Wells TL, Wyman J, et al.
Outcome measures for research in adult women with symptoms
of lower urinary tract dysfunction. Neurourol Urodyn 1998;
17:255-62.
21. Wall LL, Norton P, Versi E, Bump R. Evaluating the outcome of
surgery for pelvic organ prolapse. Am J Obstet Gynecol 1998;
178:877-9.
22. Urinary Incontinence Guideline Panel. Urinary incontinence in
adults: clinical practice guideline. Rockville (MD): Agency for
Health Care Policy and Research, Public Health Service, US De-
partment of Health and Human Services; March 1992. AHCPR
Publication No.: 92-0038.
23. Kernan WN, Viscoli CM, Brass LM, Broderick JP, Brott T, Feld-
mann E, et al. Phenylpropanolamine and the risk of hemor-
rhagic stroke. N Engl J Med 2000;343:1826-32.
24. Appell RA. Clinical efficacy and safety of tolterodine in the treat-
ment of overactive bladder: a pooled analysis. Urology 1997;
50(6A Suppl):90-6.
25. Appell RA, Sand P, Dmochowski R, Anderson R, Zinner N, Lama D,
et al. Prospective randomized controlled trial of extended-re-
lease oxybutynin chloride and tolterodine tartrate in the treat-
ment of overactive bladder: results of the OBJECT study. Mayo
Clin Proc 2001;76:358-63.
26. Gleason DM, Susset J, White C, Munoz DR, Sand PK, and the
Ditropan XL Study Group. Evaluation of a new once-daily for-
mulation of oxybutynin for the treatment of urinary urge incon-
tinence. Urology 1999;54:420-3.
Norton et al 47
Appendix
The Duloxetine Urinary Incontinence Study Group
included the following principal investigators and their
staffs: Joseph S. Sanfilippo, MD, Allegheny General Hos-
pital, Pittsburgh, Pa; Hillary Browne, MD, Alpine Clini-
cal Research Center, Boulder, Colo; Ronald L. Young,
MD, Baylor College of Medicine, Houston, Tex; Harvey
E. Armel, MD, Clinical Research Consultants, Inc, Trum-
bull, Conn; Guillermo Davila, MD, Colorado Gynecology
and Continence Center, Denver, Colo; Simon Chung,
MD, Dominion Urological Associates, Fairfax, Va; Jay M.
Baker, MD, Eastern Virginia Medical School, Norfolk,
Va; Mickey M. Karram, MD, Good Samaritan Hospital,
Cincinnati, Ohio; Ronald Tutrone, MD, Greater Balti-
more Medical Center, Baltimore, Md; Edward J. Durbin,
MD, PhD, Health Advance Institute, South Bend, Ind;
Warren O. Kessler, MD, Hillcrest Urological Medical
Group, San Diego, Calif; Veronica T. Mallett, MD, Hutzel
Hospital, Detroit, Mich; Christopher Carey, MD, Milton
S. Hershey Medical Center, Hershey, Pa; Lewis F. Russell,
Jr, MD, PA, Institute of Clinical Research, San Antonio,
Tex; Vincent Lucente, MD, Lehigh Valley Hospital, Al-
lentown, Pa; Martin Conway, MD, Lovelace Scientific Re-
sources, Inc, Albuquerque, NM; Aaron Kirkemo, MD,
Metropolitan Urologic Specialists, Detroit, Mich; James
Schoeck, MD, Orlando Clinical Research Center, Or-
lando, Fla; Alan D. Garely, MD, North Shore University
Hospital, Great Neck, NY; Karny Jacoby, MD, Northwest
Outpatient Medical Center, Seattle, Wash; Enrique M.
deCastro, MD, Northwest Women’s Clinic, Portland,
Ore; Roger D. Fincher, MD, PC3 Inc, Spokane, Wash;
Brenda Williams, MD, PC3 Inc, Boise, Idaho; John P.
Lenihan, Jr, MD, Pacific Coast Clinical Coordinators,
Tacoma, Wash; Thomas Melchione, MD, Pacific Coast
Clinical Coordinators, Sacramento, Calif; Stuart A.
Sarshik, MD, Pennridge Urological Associates PC, Sell-
ersville, Pa; David C. Reed, MD, Pacific Coast Clinical Co-
ordinators, Bellevue, Wash; Mark M. Blatter, MD,
Primary Physicians Research, Pittsburgh, Pa; Jaroslava
Zoubek, MD, Portland Clinic, Portland, Ore; Martha C.
Storrie, MD, Research Across America, Dallas, Tex; Wulf
H. Utian, MD, PhD, Rapid Medical Research, Cleveland,
Ohio; David F. Mobley, MD, Research for Health, Inc,
Houston, Tex; Linda Brubaker, MD, Rush Presbyterian-
St Luke’s Medical Center, Chicago, Ill; William E.
Friedel, MD, San Diego Urology Center, San Diego,
Calif; Kandasamychetty Perumal, MD, Seton Health In-
continence Center, Troy, NY; Don I. Boychuk, MD, Sharp
Rees-Stealy Medical Center, San Diego, Calif; Gamal
Ghoniem, MD, Tulane Medical School, New Orleans, La;
David Ellis, MD, URO Rehab, Bryn Mawr, Pa; Larry I. Gil-
derman, MD, University Clinical Research Associates,
Inc, Pembroke Pines, Fla; Sandra Culbertson, MD, Uni-
versity of Chicago Hospital, Chicago, Ill; Ellen Wells,
48 Norton et al
MD, University of North Carolina Hospitals, Chapel Hill,
NC; Peggy A. Norton, MD, University of Utah, Salt Lake
City, Utah; Anne L. Viselli, MD, University of Vermont
Women’s Center, Williston, Vt; J. Thomas Benson, MD,
Urogynecology Associates, Indianapolis, Ind; John R.
July 2002
Am J Obstet Gynecol
Miklos, MD, Urogynecology, PC, Alpharetta, Ga; Peter
Knapp, MD, Urology of Indiana, Indianapolis, Ind; Nor-
man R. Zinner, MD, Western Clinical Research, Inc, Tor-
rance, Calif; and Ananias Diokno, MD, William
Beaumont Hospital, Royal Oak, Mich.