Non-pregnant women 20 years of age and older with predominant symptoms of SUI, were enrolled in this double-blind, randomized, parallel, placebo-controlled, clinical trial conducted at 9 study centers in Taiwan. Written informed consent was obtained from all participants and study design was reviewed by a local ethics committee in accordance with the Declaration of Helsinki. Concomitant medications including urinary continence promoting drugs, antidepressants, drugs for obesity (including over-the-counter appetite suppressants and diet pills), and illicit drugs were not allowed in the study. Enrolled women reported the predominant symptoms of SUI during the last 3 months with an average of ≥ 1 incontinent episode/day. Additional history requirements included daytime voiding frequency ≤ 8 voids daily, nocturnal frequency ≤ 2 voids daily and no predominant urge incontinence symptoms. Women unable to tolerate retrograde bladder filling to 400 mL or who had a first sensation of bladder filling at ≤ 100 mL were excluded. A positive cough stress test was required after filling the bladder.
The study design and timing of acquisition of diaries and other variables are depicted in Figure . After a 2-week, no drug, lead-in period, women were randomized to 80 mg duloxetine (40 mg twice daily) or placebo for 8 weeks with post-randomization evaluation every 4 weeks. The treatment phase of the core registration trials was 12 weeks; however, the majority of the these adverse events emerged within the first 4 weeks and the number of duloxetine responders (≥ 50% reduction in median percent incontinence episode frequency [IEF]) did not change significantly after 4 weeks post-randomization. An 8-week trial was therefore considered sufficient.
Study design and the timing of acquisition of urinary diary and quality of life measurements reported.
Randomization was controlled by a computerized interactive voice response system at a central location for all study sites. Stratified randomization using baseline IEF of <14 or ≥14 episodes/week obtained from patient diaries was used to prevent potential imbalance in incontinence severity.
Weekly paper diaries were also used to collect the number of voids, the time of voids, and the number of continence pads used. Diaries were collected to determine baseline incontinence severity the last week before visit 2 during the no drug lead-in period (Figure ).
The primary efficacy variable in this study was percent change in IEF/week from baseline to endpoint, which was calculated from subject completed, real-time, paper diaries. Approximately 50% reduction in IEF has been generally accepted as a clinically relevant threshold for response in SUI outcomes research for interventions including bladder training and pelvic floor muscle training [22
], devices [23
], surgery [24
], and a pharmacological agent [18
]. An IEF responder was therefore defined as a woman who had a > 50% decrease in IEF with treatment.
Secondary efficacy variables included: 1) Incontinence Quality of Life (I-QOL) questionnaire total and subscale scores [27
], 2) Patient Global Impression of Improvement (PGI-I) rating[28
], 3) mean time between voids/day, and 4) continence pad use/week.
The 22-item I-QOL questionnaire is a globally-validated disease-specific instrument endorsed by the International Consultation on Incontinence, which was administered at each visit. The I-QOL questionnaire evaluates the effects of UI in 3 domains: avoidance and limiting behavior, social embarrassment, and psychosocial impact. An I-QOL score of 100 represents the best possible quality of life and 0 represents the worst possible quality of life. The questionnaire has not been specifically validated in Chinese language.
The PGI-I rating is a globally-validated 1-question questionnaire and was obtained at each post-randomization visit. The PGI-I measures the improvement the subject perceives in her condition since starting treatment [28
]. The questionnaire has not been specifically validated in Chinese language.
Compliance with the required study drug regiment was examined at each visit following initiation of treatment. Unused study drug was returned to Eli Lilly and Company, and compliance was assessed by counting returned study drug. Investigators encouraged compliance with study medication but subjects were not discontinued from the study for poor compliance only.
Safety was assessed by evaluation of treatment-emergent adverse events (TEAEs), discontinuations due to adverse events, serious adverse events, discontinuation emergent adverse events (DEAE), vital signs 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 double-blind treatment period. DEAE were adverse events that occurred following the end of the treatment phase and were reported during the two week follow up.
A serious adverse event was defined according to the International Consultation on Harmonization guidelines and included any adverse events associated with death, initial or prolonged inpatient hospitalization, a life-threatening experience (ie, immediate risk of dying), persistent or significant disability/incapacity, congenital anomaly/birth defect, or is significant for any other reason.
The statistical analysis plan was specified a priori and was performed in accordance with intent-to-treat (ITT) principles. Subjects with baseline and at least 1 post-baseline measurement were included in the analysis. The percent change in IEF was compared between treatment groups using the van Elteren's test, a type of stratified Wilcoxon test, with baseline incontinence severity as the stratification variable. This primary analysis compared IEF before and after randomization, pooling all diaries between visits 1 and 2 for the baseline and all diaries between visits 2 and 4 for the end point. The changes in I-QOL scores were analyzed using an ANCOVA model that included terms for baseline scores, treatment, site, and baseline incontinence severity. The endpoint PGI-I was analyzed using the Cochran-Mantel-Haenszel test with the baseline incontinence severity as the strata. The missing endpoint values in the above analyses were imputed via LOCF.
Enrollment in the study was set to end when approximately 120 women (60 per treatment group) had been assigned to a treatment group. The sample size was determined to provide 97.3% power for detecting a treatment difference of 23% in the median percent change in IEF from baseline to endpoint.
Analyses were performed using SAS 8.1 (SAS Institute, Cary, North Carolina). A two-sided alpha level of 0.05 was considered significant for treatment effects.