Prof. Dr H Drutz, Mount Sinai Hospital, Ontario Power Generation Building, 700 University Avenue, Suite 3097, Toronto, Ontario M5G 1Z5, Canada. Email firstname.lastname@example.org
Stress urinary incontinence (SUI) is a common condition affecting millions of women worldwide. It has a significant impact on the quality of life (psychosocial, social and economic well-being) of sufferers and their families. Until recently, treatment options for SUI were limited to conservative treatments such as pelvic floor muscle training, which usually has a poor compliance over time, or surgical procedures that carry the risk of complications and are mainly an option for more severely affected women. Duloxetine, a potent and relatively balanced serotonin and noradrenaline reuptake inhibitor, has been evaluated in phase II and phase III clinical trials and was found to be efficacious and safe in the treatment of women with moderate to severe SUI symptoms. Even in a subgroup of women with severe SUI awaiting surgery, with or without the presence of intrinsic sphincter deficiency, duloxetine was found to be effective. In addition, a good correlation was found between efficacy outcome measures and ratings of global impression scores. Women begin to perceive themselves as being better with treatment when their incontinence episodes frequency (IEF) decreases by 46% or their Incontinence Quality of Life (I-QOL) score improves by 6.3 points. The improvements in IEF and I-QOL obtained with duloxetine were well above these threshold levels, whereas these with placebo were not. All duloxetine responses were observed within 2 weeks. Overall, duloxetine is an effective treatment for a wide variety of women presenting with SUI symptoms. It has been approved in Europe for the treatment of women with moderate to severe SUI symptoms and is now available in many European countries.
Stress urinary incontinence (SUI) is the complaint of involuntary leakage of urine on effort or exertion, or on sneezing or coughing.1 Although SUI is not a life-threatening condition, millions of women have SUI symptoms and are dealing with the substantial impact of these symptoms on their psychosocial, social and economic well-being.2
Urinary incontinence is almost three times as common in women compared with men. SUI is the most prevalent type of incontinence (occurring in 50% of all incontinent women), followed by mixed urinary incontinence (MUI) in 36% and urge urinary incontinence in 11%.3 Whereas pure SUI is the most prevalent type in young and middle-aged women, MUI becomes more prevalent in older women.4
The function of the lower urinary tract to store and periodically release urine is dependent on a complex innervation and the normal functioning of the detrusor muscle, urethral smooth muscle, the striated rhabdosphincter and the pelvic floor muscles.5 SUI occurs when bladder pressure exceeds urethral closure pressure, resulting in transient sphincter opening and urine loss.
Duloxetine, a relatively balanced serotonin (5-hydroxytryptamine [5-HT]) and noradrenaline (NA) reuptake inhibitor, is the first pharmaceutical agent approved by the European Medicine Evaluation Agency for the treatment of moderate to severe SUI symptoms in women. Duloxetine inhibits presynaptic 5-HT and NA reuptake, resulting in increased concentrations of these neurotransmitters in the synaptic cleft, leading to sustained stimulation of postsynaptic 5-HT and NA receptors.6,7 This effect of duloxetine is mediated centrally via stimulation of the rhabdosphincter (external urethral sphincter) through the pudendal nerve and affects urethral closure during the storage phase only.8
Due to the continence-promoting properties of duloxetine observed in animal studies,5,9 a clinical trial program was carried out to asses the potential of duloxetine as a treatment for women with SUI symptoms. The efficacy of duloxetine in this patient population has been demonstrated in four published randomised control trials (RCTs) comprising a total of 1913 women randomised to duloxetine 40 mg twice a day or placebo during 12 weeks.10–13
This paper will focus on the efficacy data available from data pooling of the four RCTs and differentiate between the initial management of SUI on the one hand and women awaiting surgery on the other hand. Furthermore, the relationship between treatment naivety, incontinence severity and placebo response will be discussed.
Pooled analysis of four RCTs
Study population and methods
Four randomised, double-blind placebo-controlled trials on duloxetine had virtually comparable study design and study populations and comprised one phase II study and three phase III studies. The phase II study included women with greater than or equal to four SUI episodes per week, whereas in the phase III studies, women with greater than or equal to seven weekly SUI leakages were included. They had to have a normal day and nocturnal voiding pattern and no urge incontinence symptoms. Furthermore, women had to have a positive cough stress test and positive stress pad test (>2 g) after bladder filling (400 ml of saline bladder filling at 100 ml per minute in supine position). The women were excluded if they had a first sensation of bladder filling of <100 ml or if they had no sensation at all during filling. After a 2-week lead-in period without treatment followed by a 2-week placebo lead-in period, women were randomised under double-blinded conditions to receive either duloxetine 40 mg twice a day (n= 958) or placebo (n= 955) for 12 weeks.
At baseline, women rated their bladder function by the Patient Global Impression of Severity (PGI-S). By means of this validated self-assessment test, a woman rates her urinary tract condition as normal, mild, moderate or severe. Efficacy parameters were the median percent change in incontinence episode frequency (IEF) obtained from the woman micturition diary, the mean change in the Incontinence Quality of Life (I-QOL) score and the Patient Global Impression of Improvement (PGI-I), i.e. percentage of women reported to be at least ‘a little better’. Analyses were based on the intent-to-treat population.
1Figure 1 shows the self-assessed baseline severity of bladder function of the study population, using the PGI-S. In the phase II trial, 43.4% of women rated their condition as moderate to severe, compared with around 70% in the phase III trials.
Pooled analysis revealed that the duloxetine-treated group had a significantly greater median decrease in IEF (52%) compared with the placebo group (33%) (P < 0.001). The median decrease in IEF could not be attributed to more frequent voiding as the average voiding interval also increased significantly (mean increase: 15–20 versus 1.7–8.5 minutes on placebo).14 Moreover, the maximum treatment effect in terms of improvement in IEF was already noticeable at the first study visit, i.e. after 4 weeks of treatment, and persisted throughout the 12-week double-blind treatment period. In addition, mean total I-QOL score was significantly improved in the duloxetine group (9.2) compared with the placebo group (5.9) (P < 0.001).15 Significantly more women in the duloxetine group stated that their condition had improved (i.e. at least a little better) than in the placebo group (PGI-I: 64.6 versus 50.1%; P < 0.001).151Table 1 summarises the impact of duloxetine on the major efficacy variables in the phase II and phase III trials.
Table 1. Effect of duloxetine on the major efficacy variables in the phase II and III trials (ITT analysis)
In a published trial,15 109 women with more severe SUI (IEF ≥ 14) (Figure 1) and on the waiting list for incontinence surgery were randomised to placebo (n= 54) or duloxetine 40 mg twice a day (n= 55) for 4 weeks, escalating to 60 mg twice a day for another 4 weeks. Scheduled surgical procedures included tension-free vaginal tapes (46.2%), retropubic urethropexies (38.7%) and slings (13.2%). Intrinsic sphincter deficiency (ISD, defined as urodynamic stress incontinence, with a maximum straining urethral axis less than 20°, maximum urethral closure pressure less than 20 cm H2O or Valsalva leak-point pressure less than 60 cm H2O16) was diagnosed in 29% of women in the duloxetine group and in 38% of women on placebo. Efficacy measurements were comparable with the efficacy measurements in the above mentioned RCTs except for the Willingness to Consider Surgery rating, an unvalidated scale specifically developed for this study. This scale included the following question: ‘Based upon your current symptoms of stress incontinence, would you consider a surgical intervention?’ Responses can vary from strongly interested to strongly not interested.
The median percent decrease in IEF was significantly greater (P < 0.001) in the duloxetine group (60%) compared with the placebo group (27%). More specifically, the absolute number of incontinence episodes was decreased (P < 0.001) in the duloxetine group (7.1/week) compared with the placebo group (2.9/week). This decrease in IEF on the one hand resulted in a significantly decreased pad usage (35 versus 5%; P= 0.008) and on the other hand, an increase in I-QOL (10.6 versus 2.4; P= 0.003). In general, 63% of the women in the duloxetine group were considered responders (showing a decrease in IEF of ≥ 50%) compared with 13.5% in the placebo group (P < 0.001). Three women on duloxetine, who did not respond to 40 mg twice a day, responded when the dose was increased to 60 mg twice a day All duloxetine responders achieved the response within 2 weeks. Duloxetine was equally effective in women with and without ISD (61.5 versus 65.6% classified as responders; P > 0.999) (2Figure 2). At the end of the study, 20% of duloxetine-treated women were no longer interested in undergoing surgery, compared with none in the placebo group.
Despite the increased incontinence severity, the efficacy of duloxetine was maintained in the more severely affected women. This subgroup showed equivalent responses to duloxetine in terms of median percent decrease in IEF as did women with moderate to severe SUI symptoms. The net improvement with duloxetine (i.e. the treatment effect of duloxetine minus the treatment effect of placebo) was even higher because of the lower placebo effect.15,17
Relationship between treatment naivety, incontinence severity and placebo response
A secondary analysis of four RCTs on duloxetine aimed at revealing the relationship between previous treatment experience (treatment naivety), incontinence severity and placebo response.17 The placebo response observed in women with SUI is sizeable but variable.17 Women reported their experience with previous continence surgery and with current pelvic floor muscle training (PFMT) at baseline. Baseline characteristics for women on duloxetine and women on placebo were very similar between the two groups; no significant differences were observed. The average IEF per week was 17. Fifty-five percent of assessable placebo-treated women suffered from severe SUI (IEF per week ≥14), 11.8% previously underwent continence surgery and 16.5% actively performed PFMT. A significantly higher (P= 0.066) placebo response was observed in less severe incontinent women when comparing with the more severe women (3Figure 3). Also, with respect to treatment naivety, in women who previously underwent continence surgery and in those actively performing PFMT, lower placebo responses were observed, although the difference was only significant for PFMT (Figure 3).
Moreover, a positive correlation between PFMT usage and placebo response within a country was found (ρ= 0.44; P= 0.0001). The IEF reduction with placebo ranged from 57% in a country with 0% PFMT usage to 21% in a country with 46% PFMT usage. In comparison, the decreases in IEF induced by duloxetine did not depend on previous treatment experience.
The placebo response, i.e. the change in a condition that occurs in response to the consumption of an inactive substance, and the acceptability of placebo-controlled clinical trials of pharmacological agents have been the subject of considerable controversy. Although it has been accepted that the placebo effect in a clinical trial with a drug is in the range of 35%, a study in 2001 challenged the existence of a true placebo effect and could not demonstrate any significant clinical effect with placebo compared with no treatment.18 It is, however, well recognised that the change in a condition observed in women on placebo reflects spontaneous remission of the condition, regression towards the mean and changes in situational factors. Furthermore, active participation in a clinical trial, behavioural therapeutic benefits (contact with investigators and staff, completion of voiding diaries) and exposure to the research setting are also part of a placebo-mediated improvement in SUI.
The observation that treatment naivety increased the placebo response by 30% to 40% can be considered an estimate for the behavioural effects of these trials. Furthermore, the observation that less severely incontinent women tended to have a greater placebo response may imply that behavioural interventions are more effective in these women.
Clinical relevance of treatment response for women with SUI
Health-related quality of life measurements translate a woman's evaluation of the impact of a disease on daily life, for example, on physical and social functioning, into a score that represents a continuous variable. However, when using such a score to quantify improvements with treatment in a disease state, it is important to evaluate both the statistical significance and the clinical relevance of an observed improvement. An analysis from two of the above mentioned RCTs10,12 including women with predominant SUI symptoms validated two global impression questionnaires (PGI-S and PGI-I) for incontinence by correlating them with several other measures of incontinence severity (IEF, I-QOL, pad test outcome).19 The data from these combined studies indicate that the PGI-S and PGI-I responses correlated well with IEF, stress pad test and I-QOL measures. Women rated their urinary tract condition as moderate or severe by means of the PGI-S scale when they had an IEF of at least 21.7 (ρ= 0.36) and an I-QOL of at least 57.9 (ρ=−0.5). Women started to feel better using the PGI-I scale when they achieved at least a decrease of 46% in the IEF or 6.3 points on the I-QOL scale (2Table 2). Pooled analyses of the four RCTs10–13 as well as the study in women awaiting surgery15 showed that the decreases in IEF and the improvements in I-QOL with duloxetine were well above these threshold levels, whereas the changes associated with placebo were not.
Table 2. Correlation between the PGI-I rating, median change in IEF and mean change in I-QOL
n= 1091 (%)
Percent change IEF
Change in I-QOL
Very much better
A little better
A little worse
Very much worse
P value between groups
Spearman's ρ (P value)
These global indexes are capable of reflecting a woman's overall appraisal of her response to treatment. It should, however, be mentioned that the construct validity demonstrated in the analysis by Yalcin and Bump19 is only applicable to a population of women with predominant SUI and cannot be generalised to men or women with other lower urinary tract symptoms. These two global impression scores (PGI-I and PGI-S) are helpful to relate commonly used quantitative measures (e.g. IEF), whose values do not translate directly into an understanding of self-perceived incontinence severity, to the patient's global impressions of their condition and treatment.
In conclusion, duloxetine is an effective treatment for a wide variety of women presenting with moderate to severe SUI symptoms. Duloxetine has a rapid onset of action. Duloxetine is also effective for women with severe SUI awaiting surgery with or without ISD. Some women who are successfully treated with duloxetine may even reconsider their willingness to proceed with surgery.
There seems to be a good correlation between the PGI-S and the PGI-I scales, the IEF and the I-QOL, which provides a valuable tool to measure the woman's global impression and appraisal of a treatment.
Conflict of interest
H.D. has received reimbursement for attending a symposium and funds for research carried out in this work for Eli Lilly and Boehringer Ingelheim.