The effectiveness of invasive and non-invasive biofeedback-assisted pelvic floor muscle training with or without electrical stimulation for the treatment of stress urinary incontinence in women: a systematic review with meta-analysis and meta-regression of randomized controlled trials

Abstract

Background:

Previous reviews evaluating biofeedback (BF)-assisted pelvic floor muscle (PFM) training (PFMT) with or without electrical stimulation (ES) for women with stress urinary incontinence (SUI) have reported conflicting results, and no consensus has been reached on the optimal parameters for this intervention.

Objectives:

To (1) evaluate the effectiveness of non-invasive and invasive BF-assisted PFMT with or without ES for SUI in women and (2) investigate whether study design characteristics and intervention parameters affect treatment effects.

Design:

Systematic review with meta-analysis and meta-regression.

Data sources and methods:

Seven English and two Chinese databases were searched from inception to May 2025. Outcomes were synthesized as standardized mean differences (SMDs) or odds ratios (ORs) with 95% confidence intervals. Meta-regression was employed to investigate the impact of study design characteristics and intervention parameters on outcomes.

Results:

This review included 21 studies involving 2373 participants. Meta-analyses revealed significant effects of invasive BF-assisted PFMT on PFM strength (SMD 0.71 (0.29, 1.14), p = 0.001), urinary incontinence (UI) symptom severity (SMD −0.33 (−0.63, −0.03), p = 0.032), quality of life (QoL; SMD −1.10 (−1.78, −0.41), p = 0.002), and improvement/cure rate (OR 2.02 (1.25, 3.26), p = 0.004). Meta-analyses also revealed significant effects of both invasive (SMD −1.23 (−1.93, −0.52), p = 0.001) and non-invasive (SMD −1.23 (−1.77, −0.68), p < 0.001) BF-assisted PFMT on urine loss severity. Meta-regression analysis revealed higher improvement/cure rates when BF-assisted PFMT involved mildly challenging exercises, tailoring to patients’ abilities, therapist supervision, sessions of <10 min, durations of >1 h/week for ⩾8 weeks, and ⩾80%–95% adherence.

Conclusion:

Invasive BF-assisted PFMT can improve PFM strength, reduce urine loss and UI symptom severity, enhance QoL, and increase the improvement/cure rate. Non-invasive BF-assisted PFMT can reduce urine loss severity. However, these results must be interpreted cautiously due to methodological flaws, risks of bias, and substantial heterogeneity across the included studies. Further high-quality studies are warranted to confirm the effectiveness of invasive or non-invasive BF-assisted PFMT with or without ES for treating SUI in women. The recommended parameters may optimize treatment effects.

Trial registration:

PROSPERO (CRD42023473798).

Introduction

The International Continence Society defines urinary incontinence (UI) as any involuntary loss of urine. ¹ The prevalence of UI varies greatly, ranging from 25% to 45% globally, ² 20% to 52% in Hong Kong, ³ and 8.7% to 69.8% in Chinese women. ⁴ However, most studies suggest that the prevalence of any UI falls between 4% and 35% among adult women. ⁵ Despite its high prevalence, only a small percentage of women experiencing UI seek medical care and receive treatment due to low expectations of treatment efficacy, embarrassment, and social stigma.^6,7

UI is categorized into stress UI (SUI), urge UI, and mixed UI (MUI), which has both stress-predominant and urge-predominant forms. ⁸ SUI is the most common type in women. It is characterized by the complaint of any involuntary loss of urine when intra-abdominal pressure increases during effort or physical exertion, or when sneezing or coughing. ⁹ In women, SUI often occurs due to weak pelvic floor muscles (PFMs) or vaginal connective tissue, leading to the insufficient urethral and bladder neck closure during increased intra-abdominal pressure, which is especially common after pregnancy and vaginal delivery.^10,11 In recent decades, there has also been an increasing number of women suffering from SUI due to aging.^5,12 SUI is associated with a heavy economic burden on individuals and healthcare systems and affects social relationships, emotional and physical health, and quality of life (QoL). ¹³ SUI also reportedly affects work function, sexual activity, and personal relationships in daily life and increases the risk of depression and anxiety in women. ¹⁴

Women with mild or moderate SUI usually undergo conservative treatments rather than surgical interventions. ¹⁵ Due to its non-invasiveness and lack of associated risks, PFM training (PFMT) is the first-line conservative treatment for SUI. ¹⁶ It reduces the impact of SUI by improving PFM strength. ¹⁷ However, the success of PFMT alone, without adjunct therapies, is often limited by the lack of training adherence due to an inability to isolate PFM contractions and poor motivation to perform the exercises. ¹⁸ Therefore, PFMT is often combined with non-invasive (e.g., a surface electrode) or invasive (e.g., a vaginal probe) biofeedback (BF) alone or in combination with electrical stimulation (ES) to facilitate accurate PFMs contraction, optimize the contraction, and retain motivation for PFMT among women with SUI. ¹⁸ However, the effectiveness of BF-assisted PFMT with or without ES remains uncertain, and studies have reported conflicting results. Previous reviews of BF-assisted PFMT with or without ES for SUI indicate that BF-assisted PFMT improves cure and improvement rates, QoL, and PFM strength.^19–21 In contrast, a review in 2019 reported that BF-assisted PFMT does not offer therapeutic benefits over alternative interventions in terms of QoL, PFM strength, and amount of urine loss. ²² However, no systematic reviews have evaluated the effectiveness of non-invasive BF-assisted PFMT for treating women with SUI. Additionally, none of the previous systematic reviews performed meta-regression to explore the impact of study design characteristics and intervention parameters (e.g., sample size, intervention week, and Physiotherapy Evidence Database Scale (PEDro) score) on treatment effects. Thus, there is a need to conduct a systematic review, meta-analysis, and meta-regression to allow more robust conclusions.

Therefore, this study aimed to (1) evaluate the effectiveness of invasive and non-invasive BF-assisted PFMT with or without ES in terms of QoL, PFM strength, severity of urine loss, improvement/cure rate, and severity of UI symptoms among women with SUI, and (2) investigate whether study design characteristics and intervention parameters impact treatment effects.

Materials and methods

This meta-analytic review was developed and reported in accordance with the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines. ²³ The PRISMA checklist is provided in Supplemental Appendix 1. This review was conducted in accordance with a protocol prospectively registered in the PROSPERO database (CRD42023473798).

Search strategy

Seven English databases (Cochrane Library, Embase, PubMed, CINAHL, PEDro, Web of Science, and Scopus) and two Chinese databases (Wanfang and CNKI) were initially searched for relevant articles from their inception to October 2023, with the searches updated in May 2025. The search strategy included a comprehensive list of medical subject headings or keywords relating to the target population (women with SUI or stress-predominant MUI), interventions (invasive or non-invasive BF-assisted PFMT with or without ES), outcome measures (severity of UI symptoms, severity of urine loss, QoL, PFM strength, and improvement/cure rate), and study design (randomized controlled trials (RCTs)). A detailed list of the search terms is provided in Supplemental Appendix 2. The reference lists of the identified articles were also manually searched to identify other relevant systematic reviews and studies.

Inclusion and exclusion criteria

Studies were included in this systematic review if they (1) were RCTs, pilot RCTs, or randomized cluster or cross-over studies; (2) included women aged over 18 years with SUI or stress-predominant MUI; (3) compared the intervention to no treatment, usual care, standard care, placebo, PFMT alone, or sham control; (4) utilized invasive or non-invasive BF-assisted PFMT with or without ES; and (5) evaluated PFM strength, QoL, improvement/cure rate, severity of urine loss, and severity of UI symptoms. Studies were excluded if they (1) were not published in English or Chinese; (2) included women with urge-predominant MUI, overactive bladder, or UI resulting from other diseases (e.g., spinal cord injury or multiple sclerosis), or included men or children with SUI; (3) were case series, case studies, single group studies, or quasi-experimental studies; (4) involved interventions such as Pilates, BF combined with behavioral therapy, Chinese herbal medications, extracorporeal magnetic stimulation, surgery or traditional Chinese medicine; or (5) were conference proceedings.

Records retrieved from the searches were imported into EndNote (version 21) for organization and management. After removing duplicate records, two authors (LZ and MUA) independently selected studies according to the selection criteria. The titles and abstracts of the retrieved references were initially screened for eligibility, followed by their full texts. Any disagreements between reviewers were resolved through discussion, and a third author (PK) was consulted when they could not be resolved through discussion.

Data extraction

Two authors (LZ and MUA) independently extracted the following data: (1) first author, country, publication year, and sample size; (2) participant characteristics (e.g., mean age); (3) intervention and control conditions; and (4) the mean and standard deviation or the number of women in the results reported at baseline (pre-intervention) and last available assessment (post-intervention).

Quality assessment

The methodological quality and risk of bias (RoB) of all included studies were evaluated using the PEDro scale and the Cochrane RoB 2 tool, respectively. The PEDro scores for the included studies were retrieved from the PEDro website (https://search.pedro.org.au/advanced-search). Two authors (LZ and MUA) independently scored the studies using the PEDro scale if they were not available in the PEDro database, and also evaluated the studies using the RoB 2 tool. The PEDro scores were categorized as follows: <4 = poor, 4–5 = fair, 6–8 = good, and 9–10 = excellent. ²⁴

The level of evidence was evaluated using the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) approach with the GRADEpro software (https://www.gradepro.org/). The GRADE classification was downgraded for RoB by one level if fewer than half of the studies included in the pooled analyses had a high RoB, and by two levels if at least half of the studies had a high RoB. ²⁵ The level of evidence was downgraded for inconsistency by one level when visual or statistical heterogeneity (I² > 50%) was present, and by two levels when both forms of heterogeneity were present. ²⁶ The GRADE classification was downgraded for indirectness by one level if there was a single or minor difference across studies in populations, interventions, comparators, or outcomes, and by two levels if there were more than two or major differences. ²⁶ The level of evidence was downgraded for imprecision by one level if the 95% confidence interval (CI) did not include a clinically important effect or if the total sample size across the studies was smaller than the estimated optimal information size (OIS) criterion, and by two levels if both were present. ²⁵ The OIS was calculated for a two-arm, parallel-group design based on data from a previous study on BF-assisted PFMT, assuming an α of 0.05 and power of 80% (β = 0.2). ²⁷ If the total sample size of the studies included in the meta-analysis was smaller than the estimated OIS, the studies providing evidence for that intervention were considered small. ²⁸ The level of evidence was downgraded for publication bias if studies were industry-funded, appeared to be sponsored by individuals who could benefit from the results, or if there were study protocols in a trial registry, but no subsequent publications. ²⁸ The overall level of evidence was evaluated as high, moderate, low, or very low. ²⁹ Two authors (LZ and MUA) independently completed the GRADE assessment for all included studies. Any disagreements were resolved through discussion, and a third author (PK) was consulted when they could not be resolved through discussion.

Data analysis

Meta-analyses were conducted using the Comprehensive Meta-Analysis software (version 3, Biostat, Inc., Englewood, NJ, USA). Studies examining similar interventions (e.g., BF-assisted PFMT with ES and BF-assisted PFMT without ES) and outcome measures were synthesized. Subgroup analyses were performed by BF type (non-invasive or invasive). Effect sizes were calculated for continuous outcomes as standardized mean differences (SMDs) with 95% CIs, and for dichotomous outcomes as odds ratios (ORs) with 95% CIs. The I² statistic was used to evaluate heterogeneity among the studies, with an I² of >50% indicating potential heterogeneity. ³⁰ A random effects model was implemented when I² was >50%, and a fixed effects model was adopted otherwise. ³⁰ Publication bias was assessed using Begg’s funnel plot and Egger’s regression asymmetry test. ³¹ A p < 0.05 was considered statistically significant.

Meta-regression analyses were performed to determine the impact of study design characteristics (e.g., sample size and PEDro score) and intervention parameters (e.g., intervention duration and adherence) on treatment effects. They were only performed when at least 10 studies reported data on the same variable. ³² The degree of challenge was determined based on the exercise position: a supine or lithotomy position was considered mildly challenging, a standing or upright position was considered moderately challenging, and the use of two or more positions was considered highly challenging. Between-study heterogeneity was quantified using the I² statistic, with a value of >50% indicating substantial heterogeneity. A p < 0.05 was considered statistically significant, and a regression coefficient (RC) between 0 and 1 was considered significant ³³ ; a larger standardized RC indicates a stronger effect size. ³⁴

Results

Study selection

The searches identified 5094 potentially relevant studies. After the manual and automatic removal of duplicates, 3887 studies were eligible for inclusion in this review. After title and abstract screening, 3789 studies were excluded, and the full texts of the remaining 98 studies were further evaluated for eligibility. Of these 98 studies, 21 met the eligibility criteria and were included in this systematic review and meta-analysis, of which 12 were included in the meta-regression analysis. The study search and selection process is illustrated in a PRISMA flowchart (Figure 1).

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart of studies through the review.

Study and patient characteristics

The characteristics of the 21 included studies are summarized in Table 1. Of the 21 included studies, 19 were prospective RCTs^27,35–52 and two were pilot RCTs.^53,54 Six were conducted in Asia,^{27,38,43,44,52,54} two in the Middle East,^41,47 seven in Europe,^{36,37,39,40,48,50,53} and six in the Americas.^{35,42,45,46,49,51} Their sample sizes ranged from 21 to 600, comprising a total of 2373 participants (1154 in the experimental group and 1219 in the control group). The mean age of the participants ranged from 42.11 to 59.3 years.

Table 1.

Characteristics of included studies (n = 21).

Author Published year Country PEDro score	Age years Sample size	Intervention and parameters	BF models	Domain and outcome measure(s)	Results
Schmidt et al. 2009 Brazil 6	Exp: 54.7 ± 6.94 (n = 10) Con: 52.09 ± 13.78 (n = 11)	Exp: Invasive BF-assisted PFMT - 2 s rapid contraction/4 s relaxation, 4 s slow contraction /4 s relaxation. - 3 times/day × 12 weeks. Con: PFMT alone - 2 s rapid contraction/4 s relaxation, 4 s slow contraction/4 s relaxation. - 3 times/day × 12 weeks.	Device developed for the study	PFM strength using perineometry Quality of life using KHQ	Pre Exp: 33.54 ± 27.43 Con: 35.29 ± 24 Post Exp: 51.12 ± 28.69 Con: 48.88 ± 19.25 Pre Exp: 63.5 ± 16.55 Con: 62.4 ± 18.85 Post Exp: 41.12 ± 15.44 Con: 49.3 ± 24.96
Aksac et al. 2003 Turkey 5	Exp: 51.6 ± 5.8 a (n = 20) Con 1: 52.5 ± 7.9 a (n = 20) Con 2: 54.7 ± 7.8 a (n = 10)	Exp: Invasive BF-assisted PFMT - 10 s contraction/20 s relaxation, 40 cycles/session, 20 min/session, 3 sessions/week for 8 weeks. Con 1: PFMT alone (Digital palpation) - 5 s contraction/10 s relaxation, 10 times/session, 3 sessions/day for first 2 weeks. - 10 s contraction/20 s relaxation, 10 times/session, 3 sessions/day for the last 6 weeks. Con 2: Control group - Did not have any exercise. Just hormone replacement therapy (estradiol hemihydrate 2 mg/day and norethisterone acetate 1 mg/day).	Myomed-932 device	PFM strength using digital palpation Severity of urine loss using the 1-h pad test PFM strength using perineometry Number of women reporting improvement/cure continence	Pre Exp: 3.3 ± 0.4 a Con 1: 3.5 ± 0.5 a Con 2: 3.3 ± 0.4 a Post Exp: 4.9 ± 0.2 a Con 1: 4.8 ± 0.4 a Con 2: 3.3 ± 0.6 a Pre Exp: 20.5 ± 1.7 a Con 1: 19.9 ± 2.5 a Con 2: 29.1 ± 3.2 a Post Exp: 1.2 ± 0.2 a Con 1: 2.1 ± 0.4 a Con 2: 28.2 ± 3.7 a Pre Exp: 19.1 ± 4.8 a Con 1: 20.3 ± 6.2 a Con 2: 18.7 ± 4.9 a Post Exp: 50.0 ± 11.5 a Con 1: 37.5 ± 8.7 a Con 2: 20.0 ± 3.9 a Exp: 16 b Con 1: 15 b Con 2: 0
Glavind et al. 1996 Denmark 6	Exp: NR (n = 19) Con: NR (n = 15)	Exp: Invasive BF-assisted PFMT - Sustained 5−10 s contraction/each position, 10 times/each position, weekly performed, a total of 4 lessons. Con: PFMT alone - Sustained 5−10 s contraction/each position, 10 times/each position, at least 3 times/day.	Dantec 21L20, Skovlunde	Number of women reporting improvement/cure continence Severity of urine loss using the 1-h pad test	Exp: 11 b Con: 3 b Pre Exp: 9 (5, 22) d Con: 12.8 (9, 44) d Post Exp: 0.8 (0, 4) d Con: 10 (2, 27) d
Pages et al. 2001 Germany 4	Exp: NR (n = 13) Con: NR (n = 27)	Exp: Invasive BF-assisted PFMT - 90 min BF introduction session. - 15 min × 5 times/week for 4 weeks. - 4 training units with 10 repetitions, 5 times/day for 2 months. Con: PFMT alone - 90 min introductory group session. -Group therapy in clinic: 5 times/week for 4 weeks, 60 min/time. - Home exercise: contract 100 times/day, specific PFMT 10 min twice/day for 2 months.	Gemini 2000	PFM strength using perineometry Number of women reporting improvement/cure continence	Pre Exp: 25 ± 8 a Con: 25 ± 15 a Post Exp: 73 ± 20 a Con: 36 ± 25 a Exp: 12 b Con: 26 b
Özlü et al. 2017 Turkey 6	Exp 1: 42.33 ± 9.66 (n = 17) Exp 2: 42.11 ± 8.33 (n = 17) Con: 42.82 ± 6.3 (n = 17)	Exp 1: Invasive BF-assisted PFMT - 10 s contraction/20 s relaxation ×40 cycles/session, 20 min/session, 3 times/week for 8 weeks. Exp 2: Non-invasive BF-assisted PFMT - 10 s contraction/20 s relaxation ×40 cycles/session, 20 min/session, 3 times/week for 8 weeks. Con: PFMT alone - First 2 weeks 5 s contraction/10 s relaxation ×5/set, 2 sets/day. - The following 2 weeks 10 s contraction/20 s relaxation ×10/set, 2 sets/day. - The following 4 weeks 10 s contraction/20 s relaxation ×10/set, 3 sets/day.	Enraf Nonius Myomed 632 or Enraf Nonius Myomed 932	Quality of life using IIQ-7 PFM strength using perineometry Number of women reporting improvement/cure continence Severity of urine loss using 1-h pad test	Pre Exp 1: 7.7 ± 5.78 Exp 2: 7.64 ± 7.18 Con: 6.7 ± 4.02 Post Exp 1: 2.64 ± 3.23 Exp 2: 2.7 ± 3.71 Con: 5.35 ± 4.24 Pre Exp 1: 36.23 ± 10.92 Exp 2: 34.88 ± 13.8 Con: 38.7 ± 10.06 Post Exp 1: 55.76 ± 11.55 Exp 2: 53.41 ± 12 Con : 44.52 ± 13.79 Exp 1: 14 b Exp 2: 14 b Con : 8 b Pre Exp 1: 11.02 ± 6.96 Exp 2: 11.02 ± 12.74 Con: 11.47 ± 11.57 Post Exp 1: 3.85 ± 4.74 Exp 2: 3.97 ± 6.59 Con: 7.7 ± 7.33
Manonai et al. 2015 Thailand 7	Exp: 46.96 ± 7.22 (n = 29) Con: 48.5 ± 6.98 (n = 32)	Exp: Invasive BF-assisted PFMT - 5 s sustained max contraction/10 s relaxation for 5−10 min; 3−5 rapid max contractions; 2 s hold contraction/4 s relaxation, 3 times/day for 16 weeks with 15 min BF session. Con: PFMT alone - 5 s sustained max contraction/10 s relaxation for 5−10 min; 3−5 rapid max contractions; 2 s hold contraction/4 s relaxation, 3 times/day for 16 weeks.	NR	Quality of life using I-QOL PFM strength using perineometry Number of women reporting improvement/cure continence	Pre Exp: 53.92 ± 18.26 Con: 51.08 ± 15.93 Post Exp: 72.57 ± 10.81 Con: 70.6 ± 15.49 Pre Exp: 23.16 ± 9.98 Con: 22.48 ± 8.43 Post Exp: 29.3 ± 13.59 Con: 30 ± 11.6 Exp: 20 b Con: 25 b
Mørkved et al. 2002 Norway 8	Exp: 47.8 ± 8.2 (n = 48) Con: 45.4 ± 8.1 (n = 46)	Exp: Invasive BF-assisted PFMT - In clinic: 6−8 s contraction/3−4 times rapid contraction. - At home: 10 high-intensity contractions/set, 3 sets/day. Con: PFMT alone - In clinic: 6−8 s contraction/3−4 times rapid contraction. - At home: 10 high-intensity contractions/set, 3 sets/day.	BF-106 BF	PFM strength using Ballon catheter Number of women reporting improvement/cure continence Severity of urine loss using the 48-h pad test	Pre-Post Exp: 12.3 (9.5, 15.1) e Con: 11.1 (8.1, 14.1) e Exp: 28 b Con: 21 b Pre Exp: 40.6 (30.2, 51.1) e Con: 44.6 (34.5, 54.6) e Post Exp: 6.5 (2.4, 10.6) e Con: 6 (3.3, 8.8) e
Bertotto et al. 2017 Brazil 6	Exp: 58.4 ± 6.8 (n = 16) Con 1: 59.3 ± 4.9 (n = 15) Con 2: 57.1 ± 5.3 (n = 14)	Exp: Invasive BF-assisted PFMT - 20 min/session, 2 sessions/week for 8 weeks. Con 1: PFMT alone - 20 min/session, 2 sessions/week for 8 sessions. Con 2: no treatment	Miotool 400 system (Miotec)	Quality of life measured by ICIQ-SF	Pre Exp: 12.7 ± 3.6 Con 1: 11.1 ± 2.9 Con 2: 11.1 ± 4.5 Post Exp: 4.5 ± 3.6 Con 1: 4.3 ± 3.2 Con 2: 10 ± 4.8
Kannan et al. 2022 Hong Kong, China 7	Exp 1: 49.3 ± 5.5 (n = 17) Exp 2: 52.5 ± 6.2 (n = 17) Con: 46.8 ± 8.3 (n = 17)	Exp 1: Non-invasive BF-assisted PFMT - Supervised: 1 session/week, 45 min/session for 4 weeks. - Unsupervised: 8 contractions/times, 3 times/day, 5 days/ week, for 24 weeks. Exp 2: Invasive BF-assisted PFMT - Supervised: 1 session/week, 45 min/session for 4 weeks. - Unsupervised: 8 contractions/3 times, 3 times/day, 5 days/week, for 24 weeks. Con: PFMT alone - Supervised: 1 session/week, 45 min/session for 4 weeks. - Unsupervised: 8 contractions/3 times, 3 times/day, 5 days/week, for 24 weeks.	Exp 1: Pelvisense Conformity Europ’eenne (CE)-approved home-use Exp 2: Kegel BF pelvic floor muscle trainer (KM518)	PFM strength using MOS UI symptoms severity using ICIQ-UI-SF Severity of urine loss using the 1-h pad test	Change Exp 1/Con: 2.53 (3.85, 1.23) f Exp 2/Con: 1.09 (3.37, 1.18) f Change Exp 1/Con: −2.47 (−4.49, −0.48) f Exp 2/Con: 1.32 (−0.77, 3.42) f Change Exp 1/Con: −11.04 (−14.0, −7.99) f Exp 2/Con: 7.21 (10.9, 3.53) f
Hagen et al. 2020 UK 7	Exp: 48.2 ± 11.6 (n = 300) Con: 47.3 ± 11.4 (n = 300)	Exp: Invasive BF-assisted PFMT -3 sets/day 16 weeks. Con: PFMT alone -3 sets/day 16 weeks.	NeuroTrac® Simplex	Number of women reporting improvement/cure continence PFM strength using MOS UI symptoms severity using ICIQ-UI-SF	Exp: 18 Con: 20 Pre Exp: 2.47 ± 0.8 Con: 2.5 ± 0.79 Post Exp: 3.31 ± 0.9 Con: 3.3 ± 0.57 Change Exp/Con: −0.09 (−0.92, 0.75) f
Berghmans et al. 1996 Netherlands 8	Exp: 46.4 ± 12.12 (n = 20) Con: 50.35 ± 10.5 (n = 20)	Exp: Invasive BF-assisted PFMT - Begin with 10 times/set (5 quick and 5 sustained), 4 sets/session, and increase 30 times/set. - 25–35 min/time, 3 times/week. Con: PFMT alone - Begin with 10 times/set (5 quick and 5 sustained), 4 sets/session, and increase 30 times/set. - 25–35 min/time, three times/week.	Vaginal probe (Verimed Inc., Coral Springs, FL) and electromyograph (Myaction 12, Uniphy BV, Son)	Number of women reporting improvement/cure continence Severity of urine loss using by 48-h pad test	Exp: 19 b Con: 17 b Pre Exp: 26.63 ± 24.47 Con: 28.98 ± 31.66 Post Exp: 12.22 ± 15.4 Con: 12.5 ± 12
Sung et al. 2000 South Korea 4	Exp: NR (n = 30) Con: NR (n = 30)	Exp: Invasive BF-assisted PFMT with ES - 20 min/session, 2 sessions/week for 6 weeks. - ES: followed by invasive BF and lasted for 32 s. Con: PFMT alone - Do every day and visit the clinic once a week for 6 weeks.	Elite compact model, ECL electromedical	PFM strength using perineometry	Pre Exp: 35.4 ± 8.5 Con: 37.7 ± 7.2 Post Exp: 41.5 ± 9.8 Con: 38.7 ± 7.8
Barnes et al. 2021 Mexico 6	Exp: 44 ± 10 (n = 27) Con: 48 ± 13 (n = 27)	Exp: Invasive BF-assisted PFMT - 2 times/day for 12 weeks with four clinic visits. Con: PFMT alone - 12 weeks with four clinic visits.	PeriCoach	Quality of life using ICIQ-SF PFM strength using MOS	Change Exp: −3.95 (−2.21, −5.7) f Con: −4.73 (−3.23, −6.25) f Pre Exp: 3 (2−3) c Con: 2 (2−3) c Post Exp: 4 (3, 5) c Con: 3 (3, 4) c
Aukee et al. 2002 Finland 6	Exp: 51.8 (35.0–61.0) g (n = 15) Con: 50.8 (31.0–69.0) g (n = 15)	Exp: Invasive BF-assisted PFMT - 20 min/day, 5 days/ week. Con: PFMT alone - 20 min/day, 5 days/week.	FemiScan, Mega-Electronics	Severity of urine loss using 24-h pad test	Pre Exp: 28.1 ± 29.4 Con: 47.1 ± 34.6 Post Exp: 19.0 ± 19.7 Con: 22.5 ± 19.6
Fitz et al. 2017 Brazil 6	Exp: 56.1 ± 10.5 (n = 35) Con: 56.6 ± 12 (n = 37)	Exp: Invasive BF-assisted PFMT - Outpatient training: 10 repetitions/set for 40 min, 2 sets/week. - Home: 10 repetitions/set × 3/day for 3 months. Con: PFMT alone - Outpatient training: 10 repetitions/set for 40 min, 2 sets/week. - Home: 10 repetitions/set × 3/day for 3 months.	Neurodyn Evolution, Ibramed	PFM strength using perineometry PFM strength using MOS Number of women reporting improvement/cure continence The severity of urine loss using a 20-min pad test	Pre Exp: 20.8 (15.9, 25.7) e Con: 21.1 (15.7, 26.5) e Post Exp: 33.3 (27.6, 39.0) e Con: 32.1 (23.3, 40.9) e Pre Exp: 2.5 (2.3, 2.7) e Con: 2.4 (2.0, 2.8) e Post Exp: 3.7 (3.4, 4.0) e Con: 3.4 (3.1, 3.7) e Exp: 23 b Con: 14 b Pre Exp: 20.4 (12.3, 28.5) e Con: 19.4 (10.4, 28.4) e Post Exp: 3.7 (0.5, 6.9) e Con: 7.2 (0.8, 13.6) e
Hirakawa et al. 2013 Japan 5	Exp: 55.3 ± 9.8 (n = 23) Con: 58.3 ± 11.2 (n = 23)	Exp: Invasive BF-assisted PFMT - 10 s sustained max contraction with 5 s hold/10 s relaxation, followed by 10 s rapid contraction/4 s relaxation for 12 weeks. Con: PFMT alone - 10 s sustained max contraction with 5 s hold/10 s relaxation, followed by 10 s rapid contraction/4 s relaxation for 12 weeks	FemiScan and Mega-Electronics	UI symptoms severity using ICIQ-SF PFM strength using perineometry Severity of urine loss using the 1-h pad test Number of women reporting improvement/cure continence	Pre Exp: 11.2 ± 3.9 Con: 12.0 ± 3.5 Post Exp: 7.8 ± 3.3 Con: 8.3 ± 3.5 Pre Exp: 20.7 ± 15.2 Con: 18.3 ± 9.0 Post Exp: 33.9 ± 17.5 Con: 29.2 ± 14.3 Pre Exp: 21.3 ± 38.2 Con: 11.7 ± 18.9 Post Exp: 9.9 ± 15.1 Con: 7.7 ± 15.4 Exp: 16 b Con: 18 b
Burns et al. 1993 US 6	Exp: NR (n = 40) Con 1: NR (n = 43) Con 2: NR (n = 40)	Exp: Invasive BF-assisted PFMT - 10 rapid and 10 sustained contractions/session. - 4 sessions/day for 8 weeks. Con 1: PFMT alone - 10 rapid and 10 sustained contractions/session. - four sessions/day for 8 weeks. Con 2: Control - No treatment.	J&J Model M-53	Number of women reporting improvement/cure continence	Exp: 27 b Con 1: 26 b Con 2: 7 b
Zhu et al. 2022 China 5	Exp: 28.4 ± 3.69 (n = 55) Con: 27.66 ± 3.5 (n = 55)	Exp: Invasive BF-assisted PFMT with ES - No less than 3 s contraction/2−6 s relaxation, 15−30 min, 3 times/day, 3 months. - ES: 10−20 min, twice/week for 5 weeks. Con: PFMT alone - No less than 3 s contraction/2−6 s relaxation, 15−30 min, 3 times/day, 3 months.	NR	Quality of life using I-QOL UI symptoms severity using ICIQ-SF Number of women reporting improvement/cure continence	Pre Exp: 61.38 ± 3.76 Con: 61.99 ± 4.13 Post Exp: 77.56 ± 3.49 Con: 70.41 ± 3.14 Pre Exp: 6.51 ± 2.67 Con: 6.89 ± 2.51 Post Exp: 1.39 ± 1.72 Con: 4.84 ± 3.02 Exp: 52 b Con: 39 b
Weinstein et al. 2022 USA 6	Exp: 52.02 ± 12.9 (n = 143) Con: 51.5 ± 12.6 (n = 156)	Exp: Invasive BF-assisted PFMT - 15 s contraction/15 s relaxation ×5 times, 3 times/day, 8 weeks. Con: PFMT alone - 15 s contraction/15 s relaxation ×5 times, 3 times/day, 8 weeks.	Eva Digital Therapeutic	Quality of life using IIQ-7 Number of women reporting improvement/cure continence	Pre Exp: 38.4 ± 25.7 Con: 40.6 ± 26.8 Post Exp: 22.2 ± 19.2 Con: 25.6 ± 22.9 Exp: 59 b Con: 37 b
Wang et al. 2024 China 6	Exp: 34 (31−36) c (n = 223) Con: 34 (31−36) c (n = 229)	Exp: Invasive BF-assisted PFMT - 6 min/set, 3 sets/day, 3 months. Con: PFMT alone - 6 min/set, 3 sets/day, 3 months.	XFT-0010CK BF	Quality of life using I-QOL UI symptoms severity using ICIQ-UI-SF PFM strength using perineometry	Pre Exp: 89 (76−102) c Con: 91 (76−101) c Post Exp: 98 (88−107) c Con: 98 (89−107) c Pre Exp: 7(4−9) c Con: 6(4−9) c Post Exp: 4 (2−5) c Con: 4 (3−6) c Pre Exp: 19 (11−29.5) c Con: 16 (9−23.5) c Post Exp: 26 (17−38) c Con: 21(13.5–33.5) c
Kołodynska et al. 2022 Poland 5	Exp: NR (n = 20) Con: NR (n = 20)	Exp: Invasive BF-assisted PFMT with ES - 5 s contraction/10 s relaxation, last 5 min, daily, 2 weeks. Con: No treatment	MyoPlus4Pro 4-channel device	Severity of urine loss using the 1-h pad test	Pre Exp: 16.6 ± 16.4 Con: 6.7 ± 1.81 Post Exp: 7.85 ± 6.35 Con: 7.15 ± 1.76

^aMedian ± standard deviation.

^bNumber of women.

^cMedian (First Quartile–Third Quartile).

^dMedian (95% confidence interval).

^eMean (95% confidence interval).

^fMean difference (95% confidence interval).

^gMean (min–max).

BF, biofeedback; Con, control group; ES, electrical stimulation; Exp, experimental group; ICIQ-SF, International Consultation on Incontinence Questionnaire-Short Form; ICIQ-UI-SF, International Consultation on Incontinence Questionnaire-Urinary Incontinence-Short Form; IIQ-7, Incontinence Impact Questionnaire, Short Form; I-QOL, Incontinence Quality of Life Questionnaire; KHQ, King’s Health Questionnaire; min(s), minute(s); MOS, Modified Oxford Scale; NR, not reported; PEDro, Physiotherapy Evidence Database; PFM, pelvic floor muscle; PFMT, pelvic floor muscle training; s, second.

The examined interventions included invasive BF-assisted PFMT (n = 18),^{27,35–37,39–49,51,53,54} invasive BF-assisted PFMT with ES (n = 3),^38,50,52 and non-invasive BF-assisted PFMT (n = 2).^47,54 The control condition was PFMT alone in all included studies (n = 21).^27,35–54 Of the 21 included studies, eight assessed QoL,^{27,42,44,45,47,49,51,52} of which three utilized the Incontinence Quality of Life Questionnaire (I-QOL),^27,44,52 two utilized the International Consultation on Incontinence Questionnaire-Short Form (ICIQ-SF),^45,49 two utilized the Incontinence Impact Questionnaire, short form (IIQ-7),^47,51 and one utilized the King’s Health Questionnaire (KHQ). ⁴² Thirteen studies assessed PFM strength using the Modified Oxford Scale (MOS) or perineometry.^{27,38–44,46–49,54} Thirteen studies reported the number of women who reported improvement/cure of incontinence.^{35–37,39–41,43,44,46–48,51,52} Ten studies assessed the severity of urine loss using a pad test.^{36,37,40,41,43,46,47,50,53,54} Five studies assessed the severity of UI symptoms using the International Consultation on Incontinence Questionnaire-Urinary Incontinence-Short Form (ICIQ-UI-SF).^{27,43,48,52,54}

Quality assessment

The RoB of the included studies is summarized in Figure 2. Thirteen studies failed to clearly report their concealed allocation procedure, resulting in an unclear or high risk rating of the randomization process.^{27,35,38,39,42–44,47,50–54} Ten studies were found to have an unclear or high RoB in the outcome measurement due to a lack of assessor blinding.^{38,39,41,43,45,46,49–52} Sixteen studies presented unclear RoB of the selective reporting bias.^{27,35–39,41–44,46,47,49,50,52,53} Overall, of the 21 included studies, 2 had a low RoB,^40,48 12 had a moderate RoB,^{27,36,37,42–44,47,50–54} and 7 had a high RoB.^{35,38,39,41,45,46,49}

Figure 2.

Summary of findings (Cochrane RoB 2).

The PEDro scores of the included studies are presented in Table 1, and the PEDro assessment of the included studies is reported in Supplemental Appendix 3. The mean PEDro score of the 21 included studies was 5.71 out of 10. Among the 21 included studies, 12 lacked assessor blinding,^{37–39,41,43,45,46,49–53} 13 did not report allocation concealment,^{27,35,38,39,42–44,47,50–54} 11 lacked intention-to-treat analysis,^{27,35,37–39,41,42,45,47,50,52} and 7 reported inadequate follow-up.^{38,39,41,43,46,48,49}

The evidence profiles for each outcome measure generated by the GRADEpro Guideline Development Tool (GDT) are summarized in Table 2. The OIS was estimated at 400 participants. The overall GRADE level of evidence for the 21 studies included in the meta-analyses ranged from very low to moderate. For QoL, the GRADE quality was very low for the studies that used the ICIQ-SF. For PFM strength, the GRADE quality was very low for the studies that utilized the MOS and low for those that utilized perineometry. For the severity of UI symptoms, the GRADE quality was moderate for the studies that utilized the ICIQ-UI-SF. For the improvement/cure rate, the GRADE quality was low for the studies that reported the number of women who reported improvement/cure of incontinence. For the severity of urine loss, the GRADE quality ranged from low to very low for the studies that utilized the pad test. Serious RoB and inconsistency were the main factors leading to the downgraded level of evidence. Regarding the ORs, the visual inspection of funnel plots and Egger’s test (p = 0.27) provided no evidence of significant publication bias across the included studies (Supplemental Appendix 4).

Table 2.

Summary of findings (GRADE).

No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	No of patients		Effect		Quality
							Intervention	Control	Relative (95% CI)	Absolute
Quality of Life (assessed with ICIQ-SF) Invasive BF-assisted PFMT vs control
245,49	Randomized trials	Very serious a	Serious b	Serious c	Serious d	None	43	56	—	SMD 1.095 SD lower (1.783 lower to 0.407 lower)	⨁◯◯◯ Very low a,b,c,d
Quality of Life (assessed with IIQ-7) Invasive BF-assisted PFMT vs control
247,51	Randomized trials	Not serious	Very serious e	Not serious	Very serious f	None	160	173	—	SMD 0.457 SD lower (1.357 lower to 0.444 higher)	⨁◯◯◯ Very lowe,f
Quality of Life (assessed with I-QoL) Invasive BF-assisted PFMT vs control
227,44	Randomized trials	Not serious	Not serious	Not serious	Serious g	None	252	261	—	SMD 0.012 SD higher (0.162 lower to 0.185 higher)	⨁⨁⨁◯ Moderate g
Muscle strength (assessed with perineometry) Invasive BF-assisted PFMT vs control
927,39–44,46,47	Randomized trials	Serious h	Serious i	Not serious	Not serious	None	418	452	—	SMD 0.714 SD higher (0.290 higher to 1.139 higher)	⨁⨁◯◯ Lowh,i
Muscle strength (assessed with Modified Oxford Scale) Invasive BF-assisted PFMT vs control
446,48,49,54	Randomized trials	Very serious j	Not serious	Not serious	Serious k	None	379	381	—	SMD 0.109 SD higher (0.035 lower to 0.252 higher)	⨁◯◯◯ Very lowj,k
Severity of urinary symptoms (assessed with ICIQ-UI-SF) Invasive BF-assisted PFMT vs control
427,43,48,54	Randomized trials	Not serious	Serious l	Not serious	Not serious	None	563	569	—	SMD 0.328 SD lower (0.628 lower to 0.028 lower)	⨁⨁⨁◯ Moderate l
Improvement/cure rate (assessed with the number of women reporting improvement/cure incontinence) Invasive BF-assisted PFMT vs control
1235–37,39–41,43,44,46–48,51	Randomized trials	Serious m	Serious n	Not serious	Not serious	None	274/696 (39.4%)	247/756 (32.7%)	OR 2.018 (1.251 to 3.256)	17 more per 100 (from 5 more to 29 more)	⨁⨁◯◯ Lowm,n
Severity of urine loss (assessed with pad test) Invasive-assisted BF PFMT vs control
936,37,40,41,43,46,47,53,54	Randomized trials	Serious o	Serious p	Serious q	Not serious	None	214	220	—	SMD 1.226 SD lower (1.929 lower to 0.522 lower)	⨁◯◯◯ Very lowo,p,q
Severity of urine loss (assessed with pad test) Non-invasive BF-assisted PFMT vs control
247,54	Randomized trials	Not serious	Not serious	Not serious	Very serious r	None	34	34	—	SMD 1.226 SD lower (1.771 lower to 0.681 higher)	⨁⨁◯◯ Low r

^aDowngraded two levels due to the overall risk of bias being “high” in half or more than half of the included studies.

^bDowngraded one level due to the substantial statistical I² (65.101%).

^cDowngraded one level due to one difference in PICO/setting.

^dDowngraded one level due to the small sample size (less than 400).

^eDowngraded two levels due to the substantial statistical and visual I² (83.119%).

^fDowngraded two levels as the sample size is small (less than 400) and the CI is clinically unimportant.

^gDowngraded one level due to a clinically unimportant CI.

^hDowngraded one level due to the overall risk of bias being “high” in at least one study, but less than half of the included studies.

ⁱDowngraded one level due to the substantial statistical I² (85.167%).

^jDowngraded two levels due to the overall risk of bias being “high” in half or more than half of the included studies.

^kDowngraded one level due to a clinically unimportant CI.

^lDowngraded one level due to the substantial statistical I² (74.825%).

^mDowngraded one level due to the overall risk of bias being “high” in at least one study, but less than half of the included studies.

ⁿDowngraded one level due to the substantial statistical I² (59.613%).

^oDowngraded one level due to the overall risk of bias being “high” in at least one study, but less than half of the included studies.

^pDowngraded one level due to the substantial statistical I² (91.087%).

^qDowngraded one level due to one difference in PICO/setting.

^rDowngraded two levels as the sample size is small (less than 400) and the CI is clinically unimportant.

BF, biofeedback; CI, confidence interval; GRADE, Grading of Recommendations, Assessment, Development, and Evaluation; ICIQ-SF, International Consultation on Incontinence Questionnaire-Short Form; ICIQ-UI-SF, International Consultation on Incontinence Questionnaire-Urinary Incontinence-Short Form; IIQ-7, Incontinence Impact Questionnaire, Short Form; I-QOL, Incontinence Quality of Life Questionnaire; OR, odds ratio; PFMT, pelvic floor muscle training; SD, standard deviation; PICO, population, intervention, comparison, outcome; SMD, standardized mean difference.

Effectiveness of interventions

Quality of life

The pooled analysis of data from two studies^45,49 of good methodological quality, very low GRADE quality, and a high RoB revealed a significant effect of invasive BF-assisted PFMT compared to the control condition on QoL measured using the ICIQ-SF (SMD −1.10, 95% CI (−1.78, −0.41), p = 0.002, n = 99; Figure 3(a)). Meta-analysis of data from two studies^47,51 of good methodological quality, very low GRADE quality, and moderate RoB revealed a nonsignificant effect of invasive BF-assisted PFMT compared to the control condition on QoL measured using the IIQ-7 (SMD −0.46, 95% CI (−1.36, 0.44), p = 0.32, n = 350; Figure A in Supplemental Appendix 5). Meta-analysis of data from two studies^27,44 of good methodological quality, moderate GRADE quality, and moderate RoB also showed a nonsignificant effect on invasive BF-assisted PFMT compared to the control condition on QoL measured using the I-QoL (SMD 0.01, 95% CI (−0.16, 0.19), p = 0.893, n = 513; Figure B in Supplemental Appendix 5).

Figure 3.

(a) Effect of invasive BF-assisted PFMT compared to control conditions on QoL using ICIQ-SF. (b) Effect of invasive BF-assisted PFMT compared to control conditions on MS using perineometry. (c) Effect of invasive BF-assisted PFMT compared to control conditions on the severity of UI symptoms using ICIQ-UI-SF. (d) Effect of invasive BF-assisted PFMT compared to control conditions on the improvement/cure rate. (e). Effect of invasive BF-assisted PFMT compared to control conditions on the severity of urine loss using the pad test. (f) Effect of non-invasive BF-assisted PFMT compared to control conditions on the severity of urine loss using the pad test.

Studies marked with A and B represent different independent groups within the same study.

BF, biofeedback; ICIQ-SF, International Consultation on Incontinence Questionnaire-Short Form; ICIQ-UI-SF, International Consultation on Incontinence Questionnaire-Urinary Incontinence-Short Form; MS, muscle strength; PFMT, pelvic floor muscle training; QoL, quality of life; UI, urinary incontinence.

PFM strength

Meta-analysis of data from nine studies^{27,39–44,46,47} of fair-to-good methodological quality, low GRADE quality, and low-to-high RoB showed a significant effect of BF-assisted PFMT to a control condition on PFM strength measured using perineometry (SMD 0.74, 95% CI (0.33, 1.14), p < 0.001, n = 887; Figure C in Supplemental Appendix 5). A subgroup analysis of these nine studies revealed a significant effect of invasive BF-assisted PFMT compared to the control condition on PFM strength (SMD 0.71, 95% CI (0.29, 1.14), p = 0.001, n = 887; Figure 3(b)).

Meta-analysis of data from four studies^46,48,49,54 of good methodological quality, very low GRADE quality, and low-to-high RoB revealed a significant effect of BF-assisted PFMT compared to the control condition on PFM strength (SMD 0.14, 95% CI (0.001, 0.282), p = 0.049, n = 827; Figure D in Supplemental Appendix 5). These studies used the MOS grading to assess abnormalities in PFM strength, with a MOS grade of III or below considered a weak or absent contraction. ⁵⁵ A subgroup analysis of these four studies revealed a nonsignificant effect of invasive BF-assisted PFMT compared to the control condition on PFM strength, as measured using the MOS (SMD 0.11, 95% CI (−0.04, 0.25), p = 0.138, n = 810; Figure E in Supplemental Appendix 5).

Severity of UI symptoms

Meta-analysis of data from four studies^27,43,48,54 of fair-to-good methodological quality, moderate GRADE quality, and low-to-moderate RoB revealed a significant effect of BF-assisted PFMT compared to the control condition on the severity of UI symptoms, as measured using the ICIQ-UI-SF (SMD −0.52, 95% CI (−0.89, −0.15), p = 0.006, n = 1149; Figure F in Supplemental Appendix 5). A Subgroup analysis of these four studies revealed a significant effect of invasive BF-assisted PFMT compared to the control condition on the severity of UI symptoms, as measured using the ICIQ-UI-SF (SMD −0.33, 95% CI (−0.63, −0.03), p = 0.032, n = 1132; Figure 3(c)).

Improvement/cure rate

Meta-analysis of data from 12 studies^{35–37,39–41,43,44,46–48,51} of fair-to-good methodological quality, low GRADE quality, and low-to-high RoB revealed a significantly greater number of women reporting improvement/cure of incontinence in the BF-assisted PFMT group than in the control group (OR 2.12, 95% CI (1.33, 3.38), p = 0.002, n = 1510; Figure G in Supplemental Appendix 5). A subgroup analysis of these studies revealed a significant effect of invasive BF-assisted PFMT compared to the control condition on the improvement/cure rate (OR 2.02, 95% CI (1.25, 3.26), p = 0.004, n = 1476; Figure 3(c)).

Severity of urine loss

Meta-analysis of data from nine studies^{36,37,40,41,43,46,47,53,54} of fair-to-good methodological quality, very low GRADE quality, and low-to-high RoB revealed a significant effect of BF-assisted PFMT compared to the control condition on the severity of urine loss, as measured using the pad test (SMD −1.27, 95% CI (−1.92, −0.63), p < 0.001, n = 468; Figure H in Supplemental Appendix 5). A subgroup analysis of these studies revealed a significant effect of invasive BF-assisted PFMT compared to the control condition on the severity of urine loss, as measured using the pad test (SMD −1.23, 95% CI (−1.93, −0.52), p = 0.001, n = 434; Figure 3(e)). The subgroup analysis also revealed that two studies^47,54 with good methodological quality, low GRADE quality, and moderate RoB also revealed a significant effect of non-invasive BF-assisted PFMT compared to the control condition on the severity of urine loss, as measured using the pad test (SMD −1.23, 95% CI (−1.77, −0.68), p < 0.001, n = 68; Figure 3(f)).

Analysis of data for 51 participants from a single study ⁴⁷ evaluating the effect of non-invasive BF-assisted PFMT compared to a control condition revealed significant beneficial effects of the intervention on QoL, as measured using the IIQ-7 (SMD −0.90, 95% CI (−1.61, −0.20), p = 0.012); PFM strength, as measured using perineometry (SMD 0.98, 95% CI (0.27, 1.70), p = 0.007); and improvement/cure rate (OR 5.53, 95% CI (1.09, 25.21), p = 0.038). Analysis of data for 51 participants from a single study ⁵⁴ evaluating the effect of non-invasive BF-assisted PFMT compared to a control condition revealed significant beneficial effects of the intervention on PFM strength, as measured using the MOS (SMD 0.94, 95% CI (0.23, 1.65), p = 0.009), and severity of UI symptoms, as measured using the ICIQ-UI-SF (SMD −1.64, 95% CI (−2.41, −0.86), p < 0.001). Analysis of data for 110 participants from a single study ⁵² evaluating the effect of invasive BF-assisted PFMT with ES revealed significant beneficial effects on the severity of UI symptoms, as measured using the ICIQ-UI-SF (SMD −1.25, 95% CI (−1.66, −0.84), p < 0.001); QoL, as measured using the I-QoL (SMD 2.34, 95% CI (1.85, 2.82), p < 0.001); and improvement/cure rate (SMD 7.11, 95% CI (1.94, 26.12), p = 0.003). Analysis of data for 60 participants from a single study ³⁸ evaluating the effect of invasive BF-assisted PFMT with ES compared to a control condition revealed a significant beneficial effect on PFM strength, as measured using perineometry (SMD 0.58, 95% CI (0.06, 1.09), p = 0.029). Analysis of data for 40 participants from a single study ⁵⁰ evaluating the effect of invasive BF-assisted PFMT with ES compared to a control condition revealed a significant beneficial effect on the severity of urine loss, as measured using the pad test (SMD −1.98, 95% CI (−2.73, −1.22), p < 0.001). Analysis of data for 21 participants from a single study ⁴² evaluating the effect of invasive BF-assisted PFMT compared to a control condition revealed a nonsignificant beneficial effect on QoL, as measured using the KHQ (SMD −0.32, 95% CI (−1.18, 0.54), p = 0.465). A detailed summary of the descriptive data is provided in Supplemental Appendix 6.

Meta-regression analysis

The meta-regression analyses included 12 studies (Table 3). They revealed that the effects of BF-assisted PFMT on the improvement/cure rate were greater when the intervention was provided for more than 1 h per week (RC 0.88, 95% CI (0.35, 1.40), p = 0.001, I² = 61%) and for at least 8 weeks (RC 0.70, 95% CI (0.15, 1.23), p = 0.01, I² = 66.23%). They also revealed that the intervention led to greater improvements in the improvement/cure rate when exercise adherence was at least 80%–95% (RC 0.58, 95% CI (0.02, 1.50), p = 0.04, I² = 3.09%), sessions lasted less than 10 min (RC 0.40, 95% CI (0.006, 0.793), p = 0.047, I² = 1.50%), and the exercises were mildly challenging (RC 0.83, 95% CI (0.06, 1.60), p = 0.04, I² = 75.37%). Moreover, they revealed that the intervention led to greater improvements in the improvement/cure rate when BF-assisted PFMT was performed with a therapist (RC 0.76, 95% CI (0.28, 1.23), p = 0.002, I² = 61%) and tailored to the participant’s ability (RC 0.85, 95% CI (0.36, 1.35), p = 0.001, I² = 61%).

Table 3.

Results of meta-regression explore the effect of study-level characteristics on the improvement/cure rate.

Variables		Interpretation
Study design
PEDro Scale	RC. (95% CI): −0.1 (−0.63, 0.42) p-value: 0.70 I2: 61% N. Study: 15	Therapist supervision in the BF-assisted PFMT for SUI in women significantly increases the improvement/cure rate.
Sample size	RC. (95% CI): −0.001 (−0.0040, 0.0009) p-value: 0.22 I2: 57.12% N. Study: 15	Tailoring BF-assisted PFMT for SUI in women to the patient’s abilities significantly increases the improvement/cure rate.
Therapist supervision	RC. (95% CI): 0.76 (0.28, 1.23) p-Value: 0.002 I2: 61% N. Study: 15
Exercise tailored to ability	RC. (95% CI): 0.85 (0.36, 1.35) p-Value: 0.001 I2: 61% N. Study: 15
Intervention features
Duration of exercise per week
<30 min/week	RC. (95% CI): −1.33 (−3.15, 0.50) p-value: 0.15 I2: 61% N. Study: 15	Requiring women with SUI to perform BF-assisted PFMT for more than 1 h per week increases the improvement/cure rate.
30 min−1 h/week	RC. (95% CI): 0.61 (−0.88, 2.20) p-value: 0.40 I2: 61% N. Study: 15
>1 h/week	RC. (95% CI): 0.88 (0.35, 1.40) p-value: 0.001 I2: 61% N. Study: 15
Intervention duration
⩾8 weeks	RC. (95% CI): 0.70 (0.15, 1.23) p-value: 0.01 I2: 66.23% N. Study: 15	Requiring women with SUI to perform BF-assisted PFMT for at least eight weeks increases the improvement/cure rate.
Follow-up duration	Coef. (95% CI): −0.01 (−0.03, 0.01) p-value: 0.28 I2: 61% N. Study: 15
Duration/session
Mild <10 min/session	RC. (95% CI): 0.40 (0.006, 0.793) p-value: 0.047 I2: 1.5% N. Study: 15	Limiting BF-assisted PFMT sessions to less than 10 min increases the improvement/cure rate of women with SUI.
Moderate 10−30 min/session	RC. (95% CI): 0.67 (−0.25, 1.59) p-value: 0.15 I2: 55% N. Study: 15
High >30 min/session	RC. (95% CI): 0.42(−1.37, 2.21) p-value: 0.65 I2: 63.46% N. Study: 15
Adherence
<80%	RC. (95% CI): −1.09 (−1.92, −0.26) p-value: 0.01 I2: 43.75% N. Study: 15	Requiring at least 80%–95% adherence to the BF-assisted PFMT increases the improvement/cure rate of women with SUI.
80%−95%	RC. (95% CI): 0.58 (0.02, 1.50) p-value: 0.04 I2: 3.09% N. Study: 15
>95%	RC. (95% CI): 0.69 (−0.17, 1.55) p-value: 0.12 I2: 15% N. Study: 15
Challenge
Mild	RC. (95% CI): 0.83 (0.06, 1.60) p-value: 0.04 I2: 75.37% N. Study: 14	BF-assisted PFMT involving mildly challenging exercises increases the improvement/cure rate of women with SUI.
Moderate	RC. (95% CI): −0.07 (−2.39, 2.24) p-value: 0.95 I2:81.33% N. Study: 14
High	RC. (95% CI): 0.76 (−0.13, 1.65) p-value: 0.09 I2: 75.02% N. Study: 14

BF, biofeedback; CI, confidence interval; h, hour; N, number; PEDro, Physiotherapy Evidence Database; PFMT, pelvic floor muscle training; RC, regression coefficient; SUI, stress urinary incontinence.

Bold indicates P < 0.05, which is statistically significant.

Discussion

The database searches identified 5094 potentially relevant records, of which 21 met the eligibility criteria and were included in this systematic review and meta-analysis. Of these 21 studies, 15 (71%) were of good methodological quality based on the PEDro quality assessment scale.^{27,35–37,40,42,44–49,51,53,54} Regarding RoB, 2 (9.5%) studies had a low RoB,^40,48 12 (57.1%) had a moderate RoB,^{27,36,37,42–44,47,50–54} and 7 (33.4%) had a high RoB.^{35,38,39,41,45,46,49} The GRADE quality of the outcome measures varied from very low to moderate. Our findings suggest that invasive BF-assisted PFMT may be advantageous in enhancing QoL and the improvement/cure rate, strengthening PFM, reducing the severity of urine loss, and alleviating UI symptoms in women with SUI compared to control conditions. They also indicate that non-invasive BF-assisted PFMT may be beneficial for reducing the severity of urine loss. Based on the results of meta-regression analyses, the following parameters are recommended for BF-assisted PFMT for women with SUI: mildly challenging exercises tailored to the patient’s ability and performed under therapist supervision, sessions lasting less than 10 min, more than 1 h per week for at least 8 weeks, and at least 80%–95% adherence.

Promisingly, invasive BF-assisted PFMT appeared to improve PFM strength, reduce the severity of urine loss, and increase the improvement/cure rate. The effect size was large for the severity of urine loss (SMD −1.23) and medium to large for PFM strength (SMD 0.71). The likelihood of reporting improvement/cure was high (OR 2.02). Despite the large effect sizes and high OR, these findings are limited by the very-low-to-low GRADE quality of the included studies. Additionally, nearly all pooled data exhibited substantial heterogeneity across the included studies, likely due to variability in the diagnostic criteria for SUI or in the devices used for BF. Therefore, the true magnitude and potentially the direction of these effects remain uncertain. ⁵⁶ Moreover, invasive BF-assisted PFMT appeared to enhance QoL^45,49 and reduce the severity of UI symptoms.^27,43,48,54 The effect size was large for QoL (SMD −1.10) and small to medium for the severity of urinary symptoms (SMD −0.33). However, these results should be treated with caution because either the meta-analysis was based on a small number of studies or the mean estimate of the effect was not large. Additionally, the results for this intervention were based on studies associated with moderate-to-high RoB, particularly in domains such as outcome measurement and allocation concealment. These limitations may have introduced systematic errors and compromised the reliability of the pooled results. Overall, the evidence for invasive BF-assisted PFMT in women with SUI remains inconclusive. Additional adequately powered and high-quality studies are warranted to confirm its effectiveness for women with SUI.

Analysis of data from the included studies revealed significant effects of invasive BF-assisted PFMT with ES on the improvement/cure rate, the severity of urine loss, QoL, PFM strength, and the severity of urine symptoms. However, these results were based on data from single studies with low-to-moderate GRADE quality. Therefore, they require replication in additional studies before conclusions can be drawn. In the studies providing evidence for the effect of BF-assisted PFMT with ES, invasive BF with ES was provided via an intravaginal probe. A Cochrane review reported that intravaginal probes caused adverse effects, including pain or discomfort and reactions to the lubricant used with the device. ⁵⁷ In addition, intravaginal probes are contraindicated for women with menopause-related vaginal dryness or atrophic vagina. ⁵⁸ Five of the included studies^{27,43,49,51,54} reported that invasive BF caused adverse reactions such as yeast infection, undefined pain and discomfort, bacterial vaginosis, vaginal irritation, and skin lacerations. Given the reported adverse events associated with intravaginal probes and their suitability for menopausal women with SUI, the clinical benefit and acceptability of this approach remain uncertain. This review found limited evidence in the included studies for the effectiveness of non-invasive BF for treating SUI in women. Therefore, we recommend that more studies evaluate the effects of non-invasive BF to further explore its potential benefits for women with SUI and its potential as an alternative to invasive BF via intravaginal probes.

Our meta-analyses revealed that non-invasive BF-assisted PFMT had superior benefits compared to the control condition in reducing the severity of urine loss, as measured using the pad test. However, these results were based on two studies^47,54 of moderate RoB and low GRADE quality, which exhibited substantial heterogeneity. Therefore, this finding should be considered cautiously. Future studies with high methodological rigor are required to confirm the effectiveness of non-invasive BF-assisted PFMT for reducing the severity of urine loss in women with SUI. Nonetheless, the effect size was large (SMD –1.23), and given the safety of non-invasive devices,^47,54 non-invasive BF-assisted PFMT may be considered in the clinical management of SUI in women.

Our findings are consistent with a recent systematic review and meta-analysis by Wu et al., which evaluated BF for treating SUI in women. ¹⁹ Their meta-analysis included 21 studies and revealed significant effects of BF-assisted PFMT on the improvement/cure rate, with a large effect size (OR 4.82). ⁵⁹ Similarly, our meta-analysis revealed a large effect (OR 2.12) for this intervention on the improvement/cure rate. Wu et al. ¹⁹ also reported significant effects for BF-assisted PFMT on QoL, as measured using the I-QOL, whereas our meta-analysis revealed a nonsignificant effect on this outcome. These differences likely reflect the number of studies included in each meta-analysis for each outcome.

Current systematic reviews and meta-analyses on this topic have failed to reach a consensus on the most effective exercise parameters for BF-assisted PFMT. ²² The treatment parameters, such as the type of contraction (fast, slow, maximum, or sustained), number of repetitions per session, frequency of sessions per week, total duration of treatment sessions, exercise position, and BF equipment utilized, varied considerably among the included studies, limiting the general applicability of our findings to clinical settings. In our study, meta-regression analyses were conducted to investigate how study design characteristics and intervention parameters impact treatment effects. Significant RCs were observed for exercise duration, treatment adherence, and exercise position in improving treatment effects, from which optimal parameters could be derived. The lack of significance for the other variables might be due to the small number of participants in the included studies, the different parameters or methods used in the BF treatment, or the failure of investigators to adhere to standard instruction guidelines for BF-assisted PFMT. Therefore, additional studies are needed to explore and standardize the optimal parameters and treatment processes for BF-assisted PFMT with or without ES.

Our study had several strengths. First, it is the first to employ meta-regression to determine the effective treatment parameters for improving the improvement/cure rate among women with SUI. Second, it employed a comprehensive search strategy that used relevant terms to identify RCTs that evaluated the effectiveness of BF-assisted PFMT with or without ES on SUI in women. Third, it employed a robust and systematic methodology to evaluate the included studies. Finally, the included studies had been conducted in numerous countries worldwide, including both middle- and high-income countries, providing widespread representation.

However, our study also had some limitations. First, the evidence level of most of the included studies, as assessed using the GRADE approach, was low, and their RoB, as assessed using the Cochrane RoB 2 tool, was moderate to high. Second, the degree of UI, differences in body mass index, menopausal status, and parity of the enrolled patients may have affected the homogeneity of the included studies. Third, the low methodological quality, high heterogeneity, and variation in the implementation of intervention protocols and BF modalities across the included studies may have affected the overall estimation of the effect, limiting the generalizability of the conclusions. Finally, only studies published in English and Chinese were considered for inclusion, which may have introduced language bias.

Conclusion

Invasive BF-assisted PFMT is beneficial for improving PFM strength, enhancing QoL, reducing the severity of urine loss and UI symptoms, and increasing the improvement/cure rate in women with SUI. However, our findings must be interpreted cautiously due to substantial heterogeneity across the included studies, as well as their fair-to-good methodological quality, very-low-to-moderate GRADE quality, and low-to-high RoB. Future, adequately powered, high-quality studies are needed to confirm the effectiveness of invasive BF-assisted PFMT for treating SUI in women, as supported by this systematic review and meta-analysis. Our findings support the beneficial effects of invasive BF-assisted PFMT with ES on PFM strength, QoL, the severity of urine loss, the improvement/cure rate, and the severity of UI symptoms. However, since these findings were based on data from single studies, they require replication in other large studies. Our findings also support the beneficial effects of non-invasive BF-assisted PFMT on the severity of urine loss in women with SUI. However, since these findings were based on a small number of studies (n = 2) with substantial heterogeneity, they must be considered cautiously. Analyses of data from single studies support the beneficial effects of non-invasive BF-assisted PFMT on QoL, PFM strength, the severity of UI symptoms, and the improvement/cure rate in women with SUI. Importantly, the evidence for the effectiveness of non-invasive BF-assisted PFMT was derived from studies with good methodological quality, low-to-moderate GRADE quality, and moderate RoB. However, more rigorous, high-quality studies that adhere to the Consolidated Standards of Reporting Trials (CONSORT) guidelines are required to provide a definitive answer on the effectiveness of non-invasive BF-assisted PFMT for SUI in women. Based on our findings, BF-assisted PFMT for women with SUI should involve mildly challenging exercises tailored to the patient’s abilities, performed under the supervision of a therapist during sessions lasting less than 10 min, for more than 1 h/week for at least 8 weeks, with a minimum of 80%–95% adherence.