Urinary incontinence (UI) is one of the most common and under-recognized health concerns among older women [1]. It can lead to serious consequences, such as social isolation, falls, fractures, and admission into long-term care. Guidelines recommend pelvic floor muscle training (PFMT) as the first-line treatment for stress or mixed UI, with adherence being crucial for achieving and maintaining treatment effectiveness.
Patients and clinicians have ranked "perception of significant benefit" as the most important facilitator of PFMT adherence [2]. However, the expected treatment response time from a PFMT program and its determinants are still unknown. Our study aimed to identify (1) when older women with UI can expect to respond to their PFMT treatment, and (2) the clinical and demographic determinants of treatment response time.

This is a secondary analysis of an assessor-blind, randomized, multi-centre trial supporting the non-inferiority of group-based compared to individual physiotherapy to treat UI in older women [3].

Population
The main trial included 362 women age 60 and over with symptoms of stress or mixed UI, a body mass index (BMI) <35kg/m2, without any other medical conditions/medication likely to interfere with treatment. Participants received an individual session with a trained pelvic floor physiotherapist to learn how to effectively contract their pelvic floor muscles (PFM) and were then randomly assigned to group-based or individual PFMT.

Intervention
All participants received the same 12 structured and progressive weekly PFMT sessions supervised by an experienced physiotherapist. The sessions consisted of 15 minutes of education and 45 minutes of exercises targeting PFM function and integration into activities of daily living. Two treatment sessions were offered in case of absence. Furthermore, women were required to perform home exercises five days per week. 

Data collection
Data from both study arms were combined as the non-inferiority hypothesis was confirmed in a previous study [3]. However, the intervention approach was still considered a potential determinant of treatment response time. 
Along with the main trial outcomes, the frequency of leakage episodes was collected using the participants’ exercise diaries and reported to the physiotherapist every week between the second and last (12th) treatment session. Leakage reduction was calculated relative to the pre-treatment 7-day bladder diary. Reductions of at least 50% were considered as the minimal clinically important difference (MCID) and reductions exceeding 70% as the incremental clinically important difference (ICID). Treatment response time was considered as the time (in number of weekly sessions) required to achieve MCID and ICID. 
Clinical and demographic potential determinants of treatment response time were selected based on known predictors of PFMT success, and/or risk factors for UI: age, BMI, parity, number of comorbidities, duration of UI symptoms, UI severity (ICIQ-UI SF), UI type (stress or mixed on the QUID), PFM strength (vaginal palpation – modified Oxford grading scheme), in addition to the intervention approach attended (individual or group-based PFMT).

Data analysis
Treatment response time was described using Kaplan–Meier Estimates. Ordinal logistic regression was used to identify the determinants of treatment response time using pre-defined treatment timeframes: after completing one quarter of the PFMT program (by session 3), half the PFMT program (between sessions 4 and 6), or within the last half of the PFMT program (between sessions 7 and 12). Effects were quantified as odds ratios (OR) with 95% confidence intervals (CI). Only participants with usable leakage report data (who attended eight or more of the 12 treatment sessions and had at least one weekly leakage report entry), were included in the analysis.

Description of the participants
From the 362 main trial participants, 337 (93%) completed at least 8 out of the 12-weekly PFMT sessions and 328 (91%) presented usable weekly leakage report data and were included in this analysis. From the included participants, missing data in the leakage report summed 25% of entries, either for participants missing a PFMT session or leakage reporting (14%) or providing qualitative rather than quantitative reports of leakage, using terms such as “less” or “more leakages” and “a lot” or “a few episodes” (11%). 
The mean age of the included participants was 68.1 years old (SD 5.9), with a median of 10 (IQR 6 to 17) leakage episodes per week at baseline. The mean treatment duration was 12.9 weeks (SD 3.1), with an overall mean attendance of 11.6 (SD 0.9). Of note, no baseline clinical or demographic imbalances were observed between participants with and without usable data, nor between those attending individual or group-based PFMT. 

Treatment response time
Median MCID (50% leakage reduction) and ICID (70% leakage reduction) were achieved by the fourth (95%CI 3.6 to 4.4) and sixth (95%CI 5.3 to 6.7) weeks of treatment, respectively, with improvements continuing until the end of the 12-week program, reaching a median 83% reduction (IQR, 67% to 96%) by session 12 (Fig 1).

Clinical and demographic determinants of treatment response time
Participants were more likely to perceive MCID (50% leakage reduction) earlier in the treatment program if they had a shorter duration of symptoms, milder UI severity, weaker PFMs at baseline, and attended individual PFM training (Fig 2A). These factors did not differ between participants that attained MCID and those that did not at the last treatment session.
Participants were more likely to perceive ICID (70% leakage reduction) earlier in the treatment program if they were younger, had a shorter duration of symptoms and milder UI severity (Fig 2B). Again, these factors did not differ between participants that reached ICID and those that did not at the last treatment session.

To our knowledge, this is the first study to assess the treatment response time from a PFMT program used to treat stress and mixed UI in women. Patients often require noticeable or measurable results to pursue treatment, and clinicians feel the need to provide evidence and feedback on potential achievable benefits. 
Our study shows that older women may experience an important reduction in leakage as early as the fourth treatment session. Understanding how the treatment program progresses can hopefully motivate patients to fully engage and assist clinicians in setting realistic treatment goals and timelines.
Furthermore, our study identified easily assessable determinants of treatment response time, including age, symptom duration, UI severity, PFM strength and the intervention approach attended. Nevertheless, all participants were as likely to gain benefits at the end of the treatment, although at varying rates.

A structured and supervised PFMT program to treat UI in older women resulted in minimal and incremental clinically important differences, which were attained by most women by the fourth and sixth weeks of treatment. Younger participants, who had a shorter duration of symptoms, milder UI severity, weaker PFM and attended individual PFMT were more likely to present earlier response to treatment. Our results strengthen the recommendation that all women with stress or mixed UI should be offered PFMT as the first-line treatment and be encouraged to fully engage and start treatment as soon as they perceive symptoms.

Epidemiology of urinary incontinence (UI) and other lower urinary tract symptoms (LUTS), pelvic organ prolapse (POP) and anal (AI) incontinence. In: Cardozo L, Rovner E, Wagg A, Wein A, Abrams P, eds. Incontinence. 7th ed. ICI-ICS International Continence Society; 2023:17-88.
Neurourol Urodyn. 2015;34(7):632-639. doi:10.1002/nau.22774
JAMA Intern Med. 2020;180(10):1284. doi:10.1001/jamainternmed.2020.2993

Continence 7S1 (2023) 101005
DOI: 10.1016/j.cont.2023.101005