Efficacy of biofeedback on stress urinary incontinence in women receiving pelvic floor muscle training:a meta-analysis

Yan-Zhen Hu,Zhi-Dan Cao,Lei Wang,Yan Wang

1Department of Graduate, Tianjin University of Traditional Chinese Medicine, Tianjin 301617, China. 2Department of Nursing,Tianjin University of Traditional Chinese Medicine,Tianjin 301617,China.

Abstract

Keywords:Biofeedback,Pelvic floor muscle training,Stress urinary incontinence,Meta-analysis

Introduction

Stress urinary incontinence (SUI) is the main type of urinary incontinence (UI), which was defined as the involuntary loss of urine during physical activity,sneezing, or coughing [1]. The prevalence of SUI is about 14.9% [2], women with SUI rarely seek medical treatment, 75% patients in Europe have never sought any treatment [3]. Most of women still view SUI as a normal physiological condition associated with the consequences of aging, and some view surgery as the only treatment available. In fact, non-surgical treatment is the first-line therapy for SUI[4].

Pelvic floor muscle training (PFMT), a form of physical therapy, is recommended by the European Urological Association as the preferred method for treating mild and severe SUI [5, 6]. High-quality evidence suggests that PFMT is an effective treatment compared to no active treatment. Biofeedback, as an auxiliary tool for PFMT, could help women exercise further and supply incentive stimulate for the cure by providing an external signal [7]. It monitors patients’physiological activities, including neuromuscular and autonomic,then amplify and transmit them as visual or audio signals to the patients [8]. There are two different ways: perineal electromyographicbiofeedback (EMG-BF) and intravaginal pressurebiofeedback (P-BF); the former is to place electrodes in the anus, while the latter is to place sensors in the vagina or rectum. Either way, signals are extracted from one’s muscles [9].Yet, the clinical effects related to the use of biofeedback remain inconclusive [9].Therefore, the objective of this meta-analysis is to examine the effectiveness of biofeedback therapy for SUI, to provide reliable evidence for clinical experience.

Methods

Inclusion and exclusion criteria

Study.RCTs in English and Chinese were searched.

Participants. Female patients aged over 18 who meet the SUI diagnostic criteria published by International Continence Society [1]. Excludes urinary tract infection,severe chronic disease,cognitive impairment,pregnancy, combined with uterine prolapse, or other muscle-related diseases.

Intervention.The intervention group used PFMT with biofeedback (including EMG-BF [12-13] and P-BF[10-12, 14-17]). In order to clarify the role of biofeedback, we excluded studies that combined biofeedback with other interventions other than PFMT;the control group used PFMT alone.

Outcome.(1)Pelvic floor muscle(PFM)strength[18]:measures the ability of PFM to produce vaginal squeeze pressure.(2)Short(up to 1 hour)pad test[19]:It’s measured in grams, the procedure is to wear a pad and drink 500 mL of water.Then start walking, up and down the stairs. Finally perform a series of intensive exercise, including sitting up, coughing, running. At the end of 20 minutes or an hour,the pad was removed and reweighted.(3)Quality of life[19]:using scales to assess the quality of life associated with SUI. (4)Social activity index [9]: measured using a visual analog gauge, indicates that women may have problems with social participation (0, impossible to participate;10,no problem to participate).

Literature search strategy

PubMed, the Cochrane Library,Wed of Science, Ovid,Ebsco, PEDro, WanFang data, VIP and CNKI databases were electronically searched to collect RCTs.The deadline for searching literature was from the establishment of the database to February 2020. The search strategy was: (“biofeedback” OR “biological feedback” AND “stress urinary incontinence” OR“SUI” AND “female” OR “women” AND “trial” OR“randomly” OR “randomized controlled trial”).Through determining the retrieval strategy repeatedly,the combination of subject and free words is used to search. And then develop a search formula based on the characteristics of each database, and consult the references included in the original literature and related reviews to determine other relevant literature. Take PubMed as an example, the literature retrieval strategy is shown in Figure 1.

Figure 1 Search strategy for studies

Literature screening and data extraction

The data of the literature was extracted independently by the two reviewers. At first, read the title and abstract of the article. When met the inclusion criteria,read the full text and selected them based on the inclusion and exclusion criteria. Disagreements were resolved by discussion and consensus with a third reviewer. The data extracted included the first author;study characteristics (i.e. year, duration, setting and design); participant characteristics (i.e. mean age,sample size and systemic therapy);measured outcomes.For studies with insufficient information,the reviewers contacted the primary authors, when possible, to acquire and verify the data.

Literature quality evaluation

According to the bias risk assessment tool provided by Cochrane Handbook for systematic reviews of interventions 5.1.0, the quality evaluation of all literature was performed by two researchers. Its contents include: ①whether the random methods is correct; ② whether the allocation is hidden; ③whether the blind methods is used for the subject and the researchers;④whether the results data is complete;⑤whether there are other sources of bias.

Statistical analysis

Meta analysis was performed by RevMan 5.3 software.First, the Chi-square was used to determine whether there was clinical heterogeneity between the studies. IfP＞0.1, I2 ＜50%, it is considered that there is no heterogeneity between studies, so a fixed effect model is selected. IfP＜0.1 and I2 ＞50%, it indicates that there is heterogeneity between studies, a random effect model is selected. Subgroup analysis can be used to determine the source of heterogeneity. If the source of heterogeneity cannot be judged, meta-analysis is performed and descriptive analysis is performed.When there is a large difference in research content, research methods,and outcome indicators from other literatures,only descriptive analysis is performed.

Numerical variable data or continuous variable data use weighted mean difference or standardized mean difference (SMD) as the effect analysis statistics, and the effect analysis calculates the 95% confidence interval(CI).

Results

Literatures search results

954 papers were initially detected, of which 845 were in English and 109 were in Chinese. 318 duplicate papers were removed by EndNote software, and 28 papers were obtained after preliminary screening of topics and abstracts. Read the full of text, 20 papers were deleted, including no relevant results, large sample loss, non-randomized controlled studies, and inability to find full texts. Finally, 8 papers were included [10-17]. The retrieval process is shown in Figure 2.

Basic characteristics and methodological quality evaluation results of inclusion study

There is a detailed description of the basic characteristics of the included study in Table 1.A total of 8 articles were included in this study, among the selected RCTs, six used a P-BF [10-11, 14-17], of which five studies expressed as cmH2O [10, 14-17]and one studies expressed as μv [11]. The other one used EMG-BF, expressed as μv [13]. the last one both used P-BF and EMG-BF[12].The basic characteristics were shown in Table1. The included studies were evaluated using the Cochrane risk of bias assessment tool. The methodological quality evaluation is shown in Figure 3.

PFM strength. The autonomic contractility of PFM was used as the main measurement index in 8 studies.7 studies [10-12, 14-17] used the pressure measurement methods in millimeters of mercury and one study [13] used the electromyographic measurement method in microvolts. In order to reduce the bias caused by different study design, one of the PFM measurement methods was excluded due to the different measurement methods and units. MD was used as the effect indicator. Heterogeneity test results showed thatP= 0.001, I2= 71%, so the random effect model was taken. Meta-analysis showed that compared with PFMT alone, biofeedback assisted PFMT can improve the self-contracting force of PFM. The difference was statistically significant (MD = 4.67,95%CI(1.86,7.49),P=0.001)(Figure 4).

Short(up to 1 hour)pad test.Four studies[10,12,15,17]took 1 hour pad test as the outcome index,and one study [14] took 20 minutes pad test as the outcome index. The methods were consistent but the measurement time was different, so the SMD was taken as the effect indicator. In addition, the random effect model was taken due to the heterogeneity test wasP＜0.01, I2= 88%. The patients who received biofeedback training has less urine leakage than the man who received PFMT alone (SMD = -1.11, 95%CI(-1.84,-0.37),P=0.003)(Figure 5).

Quality of life. Three studies [12, 13, 15] compared the quality of life of the two groups, one of them took Incontinence Impact Questionnaire as the questionnaire. Others took International Consultation Incontinence Questionnaire-ShortForm as the questionnaire. Because of the measurement unit is different, SMD is used as the effect indicator. There was a small statistical heterogeneity (P= 0.32, I2=14%)between the combined results, so the fixed effect model was taken. The results of meta-analysis showed that the scores of the trial group were lower than those of the control group, indicating that the quality of life of the trial group was better (SMD = -0.34, 95% CI(-0.67,-0.01),P=0.04)(Figure 6).

Social activity index. Three studies [10, 12, 17]compared the social activity index of two groups,all of which were measured by visual analog scale. Due the homogeneity test results showed thatP= 0.17, I2=40%, so select a fixed effect model for meta-analysis,and the results shown that the social activity index of the trial group with biofeedback was significantly better than that of the control group (MD = 0.1, 95%CI (0.06, 0.15),P＜ 0.001), which indicated that biofeedback could improve faster female social activity(Figure 7).

Table1 Basic characteristics of the included studies

Figure 2 Flow chart of included and excluded literature

Figure 3 Risk of bias summary and graph

Figure 4 The effect of PFM strength. a, comparing intravaginal P-BF assisted PFMT with PFMT alone; b,comparing perineal EMG-BF assisted PFMT with PFMT alone. PFM, pelvic floor muscle; PFMT, pelvic floor muscle training;EMG-BF,electromyographic-biofeedback;P-BF,pressure-biofeedback.

Figure 5 The effect of short (up to 1 hour) pad test. a, comparing intravaginal P-BF assisted PFMT with PFMT alone; b, comparing perineal EMG-BF assisted PFMT with PFMT alone. PFM, pelvic floor muscle; PFMT,pelvic floor muscle training;EMG-BF,electromyographic-biofeedback;P-BF,pressure-biofeedback.

Figure 6 The effect of quality of life. a, comparing intravaginal P-BF assisted PFMT with PFMT alone; b,comparing perineal EMG-BF assisted PFMT with PFMT alone. PFM, pelvic floor muscle; PFMT, pelvic floor muscle training;EMG-BF,electromyographic-biofeedback;P-BF,pressure-biofeedback.

Figure 7 The effect of social activity index. a, comparing intravaginal P-BF assisted PFMT with PFMT alone; b,comparing perineal EMG-BF assisted PFMT with PFMT alone. PFM, pelvic floor muscle; PFMT, pelvic floor muscle training;EMG-BF,electromyographic-biofeedback;P-BF,pressure-biofeedback.

Discussion

This systematic review confirms the role of biofeedback by comparing biofeedback assisted PFMT with PFMT alone. The quality of the 8 RCTs is generally not high, as two of the studies adopted the computer random method [12, 14], three studies adopted the random number table method [10, 11, 15],and all of them did not describe the specific allocation hiding method in detail. Half of the studies had shedding [12, 13-15]. Only one study was double-blind [12]. Furthermore, in this review, no consistent training methods are found, but the overall can be divided into 4 parts:find(find the PFM exactly),feel (learn how to contract), force and maintain the gains.The training time varies from 20 to 40 minutes.

The patients with SUI are usually accompanied by reduced strength of PFM contractions. In theory,training of voluntary and effective control of pelvic floor function can restore neuronal pathways and optimize peripheral target function [20]. As the result,biofeedback assisted technology could be able to regulates the central nervous system, improve the voluntary contraction of the PFM.In addition,Bo et al.[21] have concluded that biofeedback training requires proper monitoring and feedback to improve PFM and strengthen a well-designed training program. Yo et al.[22]proved that the change of mean amplitude of PFM contraction after 8 weeks of treatment was an independent predictor of whether biofeedback could improve PFM strength. The training programs in the literature of this study are similar to most studies, and the training duration is more than 8 weeks. Therefore,the conclusion of PFMT with biofeedback is positive for PFM through the meta-analysis.There will be more high-quality studies in the future to further verify the effect of biofeedback on PFM according to different training plans and training times.

Since SUI is a state of incontinence after intensive activity,it is necessary to conduct pad tests on patients.Besides,pad test is an important predictor of SUI and a reliable quantification instruments for detecting urine leakage [23, 24]. In a study by Bayrak et al. [25],biofeedback was applied to 40 patients with SUI as well as took the fully dry 12 patients and decrease in wet pads of 22 patients as remarkable indicators of outcome in their study. Due to the heterogeneity, it is impossible to combine the short (up to 1 hour) and long (from 24 hour to 48 hour) pad test, the result of this study shown that the short (up to 1 hour) pad test of the trial group with biofeedback exercise was higher than that of the control group, and there was a statistical difference. This is consistent with the research results of Babak and his colleague [26].Therefore, The clinical benefits of biofeedback may also be associated with improved perception of PFM contraction.

Quality of life analysis was the common variable for determining the effects of PFMT with biofeedback.SUI significantly affects the quality of life of patients,leading to stigma and depression [27], Robinson et al.demonstrated that the effect of UI on the quality of life of patients could be evaluated by questionnaires [28].The lower the score, the higher the quality of life. The implementation of a biofeedback program under the supervision of physical therapist, coupled with the use of visual and/or auditory signals, enable patients to perform correctly and enhance learning and understanding of pelvic floor exercises, thereby increasing patients’ satisfaction and improving treatment compliance [29]. Similarly, as satisfaction and compliance increases, so does the quality of life.The results showed that the quality of life of PFMT with biofeedback was higher than that of PFMT alone and the differences observed were significantly, this is consistent with the results of Marcos et al.[29].However, more large-sample studies are needed for follow-up investigation to determine the influence of biofeedback on the future quality of life of SUI patients.

SUI has a quite negative impact on the lives of patients. It causes discomfort, embarrassment and other emotions, leading to the social and interpersonal disruption [30]. Through the training method of accurately locating PFM assisted by the instrument could quickly and effectively improve UI, encouraging patients to go out of the house and participate in social activities [31].Sandra et al.[32]shows that the quality of life of homebound subjects is significantly lower than non-homebound subjects, which may also reflect that the increase of social activities can improve the quality of life of patients.Therefore, the social activity index of this study is consistent with the results of quality of life. Considering that the included article published far away, it is suggested that relevant scholars pay more attention to this field and conduct more high-quality research in the future.

Conclusion

In summary,the evidence that PFMT with biofeedback may be an effective treatment for SUI than PFMT alone is encouraging. The conclusion of this study can help clinicians make clearer conclusion about the patient’s plan. However, given the methodological and quality differences between RCTs, we expect more high-quality RCTs, following the minimum recommendations for clinical trails and considering the low risk of bias and adequate sample sizes, to further demonstrate its true effects.