A model identifying characteristics predictive of successful pelvic floor muscle training outcomes among women with stress urinary incontinence
Brooks K, Varette K, Harvey MA, et al. A model identifying characteristics predictive of successful pelvic floor muscle training outcomes among women with stress urinary incontinence. International Urogynecology Journal. 2020;32(3):719-728. doi:10.1007/s00192-020-04583-zLink to Original Article: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7902568/
Key Points
- The study aimed to prospectively identify factors predictive of cure with physiotherapist-supervised pelvic floor muscle training (PFMT) in women with stress urinary incontinence (SUI).
- 77 women with SUI completed the protocol, with 49% deemed cured with PFMT.
- Predictors of cure included lower severity of symptoms as measured by the International Consultation on Incontinence Questionnaire Urinary Incontinence Short Form (ICIQ-UI-SF), better bladder support in a standing position, and higher pelvic floor muscle (PFM) tone.
- Baseline bladder neck height and ICIQ-UI-SF score were significant predicting factors in the binary logistic regression model.
- The model accurately classified 74% of participants and showed 70% sensitivity and 75% specificity in predicting cure with PFMT.
- Bladder neck support and symptom severity were identified as key factors influencing the success of PFMT interventions in women with SUI.
Introduction
The research paper discusses the significant impact of stress urinary incontinence (SUI) on women, both socially and economically. It highlights that urinary incontinence (UI) affects up to 50% of women with most women having SUI and imposes a considerable financial burden on the healthcare system, costing approximately $3.84 billion CAD annually in Canada. Current interventions for SUI, such as pelvic floor muscle training (PFMT), are not always successful with only half of women being cured of SUI, and access to physiotherapy for PFMT can be challenging and costly for many women. The paper emphasizes the need for predictive models to identify factors that can streamline PFMT treatment, reduce costs, and improve the patient experience.
Previous predictive models for PFMT outcomes have been limited by inconsistent findings, use of retrospective data, and clinically inaccessible tools. Therefore, the paper aims to address these limitations by conducting a prospective study using clinically accessible measures to create a robust and practical predictive model for women's success with PFMT. The success of PFMT interventions will be evaluated objectively through the observation of a dry pad test after the intervention period. This approach aims to provide a more efficient and cost-effective means of identifying which women will benefit most from PFMT for SUI.
Methods
Participants
The study was prospective interventional cohort study. The study recruited women with stress urinary incontinence (SUI) who were ≥ 18 years old and surgical candidates for mid-urethral sling insertion from waitlists at four urogynecology clinics for a randomized controlled trial (RCT) of a physiotherapy intervention. Women with SUI were also recruited from waitlists at two physiotherapy clinics. Inclusion criteria required predominant symptoms of SUI with or without urgency incontinence or nocturia, while exclusion criteria included pregnancy or pregnancy within the previous 12 months, fecal incontinence, use of incontinence-affecting medications, neurological impairments, connective tissue disorders, detrusor overactivity, prior pelvic surgery for SUI, or pelvic organ prolapse (POP). Eligible participants were verbally instructed to complete several questionnaires, via telephone, including the International Consultation on Incontinence Questionnaire Urinary Incontinence Short Form (ICIQ-UI-SF) and Female Lower Urinary Tract Symptoms (ICIQ-FLUTS), the Medical Outcomes Survey SF-36, and a 3-day bladder diary, although the results of last three questionnaires were not used in this study. The study's focus was on women randomized to receive physiotherapy intervention from the concurrent RCT and aimed to create a predictive model for the success of pelvic floor muscle training interventions for SUI.
Baseline assessment
The baseline assessment in the research paper involved multiple steps to evaluate the participants. Prior to the assessment, the participants provided demographic and medical history information. The participants were asked to empty their bladder 1 hour prior to the assessment and drink 500ml of water. During the assessment, a 3-day bladder diary was reviewed to confirm eligibility, and underwent a trans-abdominal ultrasound to estimate bladder volume. The participants also underwent a 30 minute pad test when their bladder was estimated to be 300 to 400ml full, and the participants were excluded from the study if their pads were dry. The participants also completed physical evaluation including reflex and sensory testing of S2-5 dermatome, an evaluation of PFM strength using Modified Oxford Scale of right and left levator ani muscle, and PFM tone with a grading scale of +3 to -3.
The ultrasound analysis evaluated morphological features associated with stress urinary incontinence (SUI), such as pelvic organ support, urethral mobility, urethral morphology, and PFM function. The assessment included visualization of the entire volume of the urethra, levator hiatus, and pelvic structures at rest, during maximum voluntary contraction (MVC), maximal Valsalva maneuver (MVM), and during coughing. The participants performed specific tasks instructed by the physiotherapist, and imaging was repeated in a standing position. The assessment aimed to gather comprehensive data on pelvic floor function and urogenital morphology in relation to SUI. The thorough assessment provides detailed information for the development of a predictive model for the success of pelvic floor muscle training (PFMT) interventions for the participants with SUI.
Physiotherapy intervention
The physiotherapy intervention in the study was carried out by experienced physiotherapists with specific training in pelvic floor assessment and SUI management. The intervention comprised six face-to-face sessions conducted over a 12-week period. The initial three weeks involved weekly sessions, followed by sessions every two weeks for the next two weeks, and a final session four weeks later. The duration of the sessions ranged from approximately 30 to 45 minutes. During these sessions, the participants were instructed on proper pelvic floor muscle contractions, utilizing electromyography biofeedback in the first two sessions to facilitate motor learning. Manual therapy techniques were employed when necessary, and participants were taught to engage in a strong pelvic floor muscle contraction before activities that increase intra-abdominal pressure, known as the "knack." Additionally, a home exercise program was prescribed, involving three sets of ten maximum effort voluntary pelvic floor muscle contractions per day, with relaxation after each contraction. The exercises were progressively adjusted at each session, including changes in position and variations in contraction speed and force.
Follow-up assessment
The follow-up assessment within 2 weeks of the participants' last physiotherapy session involved the completion of a 3-day bladder diary and the ICIQ-UI-SF, as well as a return to the laboratory for a follow-up pad test, clinical assessment of pelvic floor muscle (PFM) strength and tone, and ultrasound assessment of pelvic morphology. For women participating in the randomized controlled trial (RCT), this follow-up assessment was part of the study, while for women recruited outside of the RCT, it marked the end of their participation. The same research physiotherapist who conducted the baseline assessment collected the follow-up data, remaining blind to whether participants received the physiotherapy intervention. This approach aimed to ensure consistency and minimize bias in the data collection process.
Outcome variables
The outcome variables in the research paper are focused on assessing the effectiveness of pelvic floor muscle training (PFMT) for stress urinary incontinence (SUI) in women. A successful outcome, or cure, was defined as a dry pad test result, where the pad weight gain was less than or equal to 2 grams at the follow-up assessment. Two-dimensional and 3D ultrasound imaging and video analyses were utilized to assess morphological features of pelvic organ support and urethral morphology. This included measurements such as levator hiatus area, levator plate length, bladder neck height, urethral length, and urethral wall cross-sectional area, as well as mobility assessments during coughing and maximum voluntary contraction (MVC). Pelvic floor muscle (PFM) function was measured through changes in levator hiatus area and levator plate length during MVC, as well as the cranial position of the bladder neck relative to the levator plate achieved during MVC. Additionally, levator avulsion was assessed using 3D volumes acquired during MVC, with two independent raters evaluating each volume for the presence of avulsion. In cases of disagreement between the raters, an independent reviewer intervened to reach a consensus. The outcome variables assessment employed advanced imaging techniques and detailed measurements to evaluate the impact of PFMT on SUI in women.
Statistical analysis
The statistical analysis section of the research paper outlined the methodology used to analyze the data collected from the study on pelvic floor muscle training (PFMT) interventions for women with stress urinary incontinence (SUI). Participants who did not complete the follow-up assessment were excluded from the analysis as their post-treatment outcomes were crucial for categorizing women as either cured or not cured in the model. The statistical analyses were performed using SPSS 25. The variables were initially tested for normality using the Kolmogorov-Smirnov test.
To identify potential predictors of cure in a binary logistic regression model, univariate t tests and Mann-Whitney U tests were conducted for normal and non-normal data, while Chi-squared analyses and Fisher's exact test were performed for categorical data. A significance level of α = 0.10 was set for inclusion in the predictive model. The researchers planned a priori to include only one potential predictor from each category of morphological features associated with SUI pathophysiology.
Univariate testing identified outcome variables with the largest effect size and those that did not violate the assumptions of linear regression for inclusion in the model. Missing values were handled through multiple imputations prior to generating the predictive model. The final model was subjected to goodness-of-fit testing and bootstrap validation to control for model optimism. Odds ratios were used to describe the independent associations of each predicting variable, and the diagnostic ability of the overall model was assessed using a pooled receiver operating characteristic (ROC) curve.
Overall, the statistical analysis provided a rigorous and robust methodology for creating a predictive model for the success of PFMT interventions for women with SUI, addressing key clinical and methodological considerations in the process.
Results
The study recruited 77 women with stress urinary incontinence (SUI) and found no significant differences in baseline characteristics between those who completed the study and those who did not. Univariate analyses identified 14 potential predictors for inclusion in the logistic regression model, including factors related to pelvic organ support, urethral mobility, pelvic floor muscle (PFM) function, and clinical assessment variables. Women who were cured with the PFMT intervention showed evidence of improved pelvic organ support compared to those who were not cured. This was demonstrated by higher bladder neck height in a quiet standing position, at rest in a supine position, and during MVM in a supine position. Additionally, there was shorter LPL during coughing in a supine position and during MVM in a standing position, as well as a smaller levator hiatus circumference at rest in a supine position and in a quiet standing position. These findings were also observed in the levator hiatus area. The women who were cured tended to have less urethral mobility, as evidenced by lesser bladder neck descent observed during a cough performed in a supine position and in a standing position. Participants in whom the bladder neck was positioned more cranially within the pelvis were more likely to have a dry pad test after the intervention of pelvic floor muscle training (PFMT) compared to women with lower levels of bladder neck support.
The final predictive model included bladder neck height in a quiet standing position and SUI severity, accurately classifying 74% of women as cured or not cured. The model had a Nagelkerke generalized R2 of 0.35 and an area under the curve (AUC) of 0.80, indicating effective prediction of PFMT outcomes. The logistic regression equation for predicting success with the intervention is as follows: 0.2049 (Bladder Neck Position in mm) - 0.1381 (ICIQ SF ≥ 1.0812). The logistic regression equation and estimated cut-off scores for baseline ICIQ-UI-SF score and bladder neck position provided sensitivity and specificity values for predicting PFMT success. These findings suggest that the model can effectively differentiate between women who are likely to benefit from PFMT and those who are not, with implications for individualized treatment decisions.
Discussion
The research paper discusses the development of a predictive model for the success of pelvic floor muscle training (PFMT) interventions in women with stress urinary incontinence (SUI). The model identified two significant predictors: bladder neck height in a quiet standing position and self-reported SUI severity. Women with a more cranially positioned bladder neck during standing and lower SUI severity were more likely to have successful PFMT outcomes. The study used a standardized pad test post-treatment to define cure, with a 49% cure rate. Physiotherapy was found to be more successful in women with mild to moderate SUI severity, potentially due to defects in tissue support and connective tissues. Bladder neck height was the most predictive factor, indicating better pelvic organ support as a key to successful intervention. The study also examined the role of pelvic floor muscle tone, urethral hypermobility, and levator avulsion in predicting treatment outcomes. The findings suggest that the measures included in the model are easily replicable in clinical settings. However, the study's generalizability is limited to women with specific SUI characteristics and exclusion criteria. The short-term follow-up and the need for further model validation are noted as study limitations. The paper concludes that future research should explore additional variables to enhance the model's performance and robustness before implementation in clinical practice.
Conclusions
The research paper aimed to develop a predictive model for the success of pelvic floor muscle training (PFMT) interventions for women with stress urinary incontinence (SUI). The study identified two significant predictors of PFMT success: the severity of SUI, measured by the ICIQ-UI-SF, and the bladder neck height in a quiet standing position, measured through 2D transperineal ultrasound. Women with lower symptom severity and a more cranial bladder neck position in quiet standing were more likely to demonstrate a dry pad test after the intervention. The study supports the recommendation that women with mild to moderate symptoms may be better suited to physiotherapy interventions for SUI than those with more severe symptoms, and suggests that evaluating the bladder neck position in standing may be useful for clinical decision-making. The predictive model, generated using prospectively acquired data, considered demographic, clinical, and morphological variables, and may help guide treatment recommendations for women with SUI after further revisions and validation. The study's contributions included data collection, analysis, and manuscript writing by K.C.L. Brooks and data collection and manuscript editing by K. Varette, while other team members contributed to data collection, physician partnership, manuscript editing, and project development. The study was funded by the Canadian Institutes of Health Research (CIHR) with no restrictions on publication.
Opportunities for Future Research
1. Investigate the long-term outcomes of pelvic floor muscle training (PFMT) for stress urinary incontinence (SUI) by conducting a follow-up study on the participants to assess the sustainability of the cure and identify any factors influencing long-term success.
2. Explore the use of other clinical and demographic variables, in addition to those identified in the study, to improve the predictive model for success with PFMT among women with SUI.
3. Conduct research to validate the predictive model on a new group of women with SUI to assess its generalizability and performance in diverse patient populations.
4. Investigate the impact of bladder neck support in a standing position as a predictive factor for success with PFMT, and explore its correlation with other pelvic floor morphology measures to enhance the understanding of its role in predicting treatment outcomes.
5. Explore the relevance of passive pelvic floor muscle (PFM) tissue properties, such as tone and resistance to stretch, by comparing different assessment methods, such as digital palpation and dynamometry, to determine their predictive value in the success of PFMT interventions for SUI.
6. Investigate the predictive value of other pelvic floor morphology measures, such as urethrovesical junction approximation, and their potential inclusion in the predictive model to improve the performance and robustness of the model for guiding treatment recommendations for women with SUI.
